<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" version="2.0">
  <channel>
    <title>Insights</title>
    <link>https://www.seerinteractive.com/insights</link>
    <description>Industry-leading tips, tools/techniques for optimization, analysis, and measurement. Covering SEO, Data, Paid Media, Analytics, UX, CRO, and more.</description>
    <language>en</language>
    <pubDate>Thu, 19 Nov 2026 21:04:26 GMT</pubDate>
    <dc:date>2026-11-19T21:04:26Z</dc:date>
    <dc:language>en</dc:language>
    <item>
      <title>Why Doesn’t Your Brand Appear in ChatGPT Responses? Check Its Reasoning</title>
      <link>https://www.seerinteractive.com/insights/chatgpt-reasoning</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-reasoning" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/chatgpt-reasoning.png" alt="Why Doesn’t Your Brand Appear in ChatGPT Responses? Check Its Reasoning" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;We ran 20 prompts through OpenAI's API and tracked every step and reasoning summary to understand why it picked some brands and ignored others&amp;nbsp;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;In our prompt runs, fewer than 1 in 10 brands were named before ChatGPT runs a single search (AKA around step 4 in the process)&lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;ChatGPT searches brand websites infrequently after its first search round (13% of queries) but most of the time in later rounds (75%)&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;We identified 5 main reasons a brand gets excluded: prompt-brand mismatch, brand website is missing info, known brand weakness, trust concerns, or too far down the shortlist&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt;   
&lt;h2&gt;&lt;span&gt;How Does ChatGPT Decide Which Brands Make the Cut?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Wouldn’t it be nice if we knew where ChatGPT citations came from? Or why ChatGPT searches for those specific fanout queries? For over three years now, we’ve been peeling back the layers to better understand how ChatGPT responses come together. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;First, we tracked AI visibility to understand whether your brand shows up at all. Then we focused on citation data and which brands were being sourced. After that we moved on to query fan-outs, which show how those citations are being selected.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;But many of the discussions we’ve seen talk about fan-outs as if those searches happen simultaneously. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;When you dig into ChatGPT’s reasoning, &lt;span style="font-weight: bold;"&gt;you can see a chain of thought that runs from the prompt all the way to the final response&lt;/span&gt;. Looking at the steps of this reasoning chain can help you understand which brands get mentioned, and why. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In this post, we’ll walk through how ChatGPT’s reasoning works, reasons that explain why your brand isn’t showing up, and how to fix it.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;ChatGPT’s Reasoning Is a Signal&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Reasoning is exactly what it sounds like. It’s how ChatGPT decides on the criteria for its answer: &lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;What matters for this question&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Which sources to trust&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Which brands should and shouldn’t make the list&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT’s reasoning isn’t that different from how people think. If you were buying an e-bike to commute from North Park to downtown San Diego and ride trails on the weekend, you wouldn’t start by browsing bikes. You’d think about the hills on your route, check whether e-bikes are even allowed on the trails you ride, and only then decide which features matter.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you’re a brand that sells e-bikes, ChatGPT’s reasoning can tell you a lot more than reviewing cited URLs.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT decided the shortlist of options it will mention should come down to a set of criteria, like the terrain of the commute and California’s trail rules. Understanding the reasoning helps you see what those criteria are, so you have a higher likelihood of being visible in those searches and reinforcing signals ChatGPT is searching for on your own site.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In the app, the reasoning is tucked behind “Worked for 34s” and shown in short form. That’s why we’re going to use the OpenAI API, which shows a summary of every reasoning step in the chain.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The API is directional to what actually happens in ChatGPT’s user interface. And while these results might not be 1:1 with what users are experiencing, &lt;/span&gt;&lt;strong&gt;&lt;span&gt;why not take action on the direct hints ChatGPT is providing?&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Want to see this for your&amp;nbsp;brand? &lt;/span&gt;&lt;/strong&gt;&lt;span style="font-weight: normal;"&gt;We've created a tool to capture every trace in this study. All you have to do is paste a prompt, add your OpenAI key, and watch ChatGPT reason, search, and trim in real time.&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;The Flow of ChatGPT’s Reasoning &lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;span&gt;To better understand ChatGPT’s reasoning, we ran 20 real prompts on November 4, 2026 through OpenAI’s API and read every reasoning summary ChatGPT wrote along the way. We used GPT-6 Luna at medium reasoning to align with ChatGPT’s default model. Medium reasoning was used due to its response and citation results corresponding to the app at the closest rate to other model settings tested.&lt;/span&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;span&gt;What we found is that &lt;/span&gt;&lt;strong&gt;&lt;span&gt;ChatGPT runs through several rounds of searching and reasoning before answering&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;. Let’s break down what each of those steps looks like.&lt;/span&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span style="color: #000000;"&gt;&lt;strong&gt;Step 1: The Prompt&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;This is the start of it all. ChatGPT starts its reasoning after a user submits a prompt.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;One of the prompts we looked at was about home insurance in Georgia:&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span&gt;“Which home insurance companies are most recommended for homeowners in Georgia, and why do they stand out?”&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 2: The First Reasoning Step, Where ChatGPT Builds the Plan and the Source Rules&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Depending on the prompt and your industry, ChatGPT can name those credible sources outright. &lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span style="white-space-collapse: preserve;"&gt;For our Georgia home insurance prompt, it planned to: "check credible sources, including the state insurance regulator, a ratings agency, and an insurance association."&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 3: Broad Searches&lt;/strong&gt;&lt;span style="white-space-collapse: preserve;"&gt; &lt;/span&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Next ChatGPT moves to broad searches, which include a lot of "best of" and "review" queries. These are what surface the listicles commonly discussed in AI citations, along with other articles.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span style="white-space-collapse: preserve;"&gt;For our example, its first round of searches had fan-out queries like “best homeowners insurance companies Georgia 2026 recommendations” and “Georgia homeowners insurance companies customer satisfaction [a ratings agency] 2025.”&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 4: The Second Reasoning Step, Building the Shortlist &lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;In its next reasoning summary, ChatGPT starts naming brands for consideration. In our analysis, it usually names about eight options but ends up mentioning about six in the final response.&lt;/span&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In our prompt runs, &lt;span style="font-weight: bold;"&gt;fewer than 1 in 10 are named before ChatGPT runs a single search&lt;/span&gt;. Most brands (59%) showed up right after the first round of broad searches, which is why retrieval and being visible on retrieved pages matter so much. The rest get added in later rounds.&lt;/span&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;This is also where ChatGPT adds conditions an option has to meet. About half of these reasoning notes add at least one, like “Looking into B2B SaaS routing and CRM integrations is key too.” Then it decides what it still needs to check.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span style="white-space-collapse: preserve;"&gt;After reading a personal-finance site’s Georgia guide, it wrote, “I’m considering a list of five insurance companies,” naming five insurers and flagging a sixth that “might stand out for low complaints.”&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="color: #000000;"&gt;&lt;strong&gt;Step 5: Targeted Searches&lt;/strong&gt;&lt;span style="white-space-collapse: preserve;"&gt; &lt;/span&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Once it has a shortlist, ChatGPT sends out more searches aimed at those specific brands, often using the site: operator to search just one website. In GPT-6’s first round of searching, only 13% of queries do this. In later rounds, 75% do. It finds brands on broad searches, then often goes to their own websites to check them.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span&gt;In our Georgia example, the targeted check went directly to another personal-finance site’s Georgia guide to verify its shortlist.&lt;/span&gt;&lt;/pre&gt; 
&lt;br&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 6: The Last Reasoning Step, Where ChatGPT Trims the List&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Before it writes the response, ChatGPT writes one final note where it trims the list and excludes some of the brands it found. In our prompt set, 38% of these final notes give a reason to drop or exclude an option from the set it will show the user.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span&gt;Right before writing the response, it wrote, “[An insurer] could be optional, but we might trim that,” and decided to say there’s “no universally best insurance” rather than crowning one winner.&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 7: The Response&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Based on all of its searching and reasoning, ChatGPT serves the final output that users see.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span&gt;The final response recommended the five insurers from ChatGPT's shortlist. The sixth, the one it said it “might trim,”&lt;/span&gt;&lt;strong&gt;&lt;span&gt; didn’t make it&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;.&lt;/span&gt;&lt;/pre&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;So, Why Does ChatGPT Exclude a Brand?&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p style="font-weight: normal;"&gt;After reviewing the flow from prompt to response for the prompts analyzed, we identified five primary reasons why ChatGPT excludes a brand.&lt;/p&gt; 
&lt;h3&gt;Reason 1: You Don't Fit the Exact Prompt&lt;/h3&gt; 
&lt;p&gt;Your brand isn’t a bad option, it’s just the wrong kind of option for this prompt. Your brand is built for a different buyer or use case, or you don’t serve the location the user asked about.&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT sets these conditions early. For a Baltimore HVAC prompt it wrote: “I should confirm service area and current ratings, especially for the Baltimore area.” &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt; In one case it even planned to include “two smaller local options in the quote list, even if that means leaving out a higher-ranking company.” &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;You’re not being compared against other brands at this step. You’re being evaluated against the user’s specific needs, location, and criteria. If you don’t meet them, you’re out before the comparison even starts. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Effort to fix: &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;High, often not fixable. Your brand either meets the user’s needs or it doesn’t. If the user wants a local option and you aren’t one, this isn’t a topic to prioritize.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;Reason 2: Your Site Is Missing the Signals ChatGPT Looks for&lt;/h3&gt; 
&lt;p&gt;ChatGPT goes to your official website to check you against its criteria. If the information isn’t there or isn’t easy to find, it moves on.&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In some cases, ChatGPT will name the exact information it's looking for. ChatGPT's reasoning has specifically said it’s crucial to verify credentials, check for relevant certifications, and find evidence your brand matches the criteria.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt; For the Baltimore HVAC prompt, it wrote: “Maryland requires an &lt;/span&gt;&lt;strong&gt;&lt;span&gt;HVACR license&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; for technical work and an &lt;/span&gt;&lt;strong&gt;&lt;span&gt;MHIC license&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; for home improvement contracts, so it’s crucial to verify these credentials for applicable projects.” &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt; If you’re a contractor, that means listing every license, permit, and certification that applies to your work. Some of them ChatGPT may never check for, but it won’t hurt to have them on your site, and you can’t predict every requirement it’ll look for. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;So brands should know which requirements ChatGPT is looking for. Some brands have them, but the information isn’t easy for users or AI to find, so they’re excluded by default.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Effort to fix: &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Low, this is the one to focus on. This could be the most tactical way to improve your AI performance: surface information you already have that just isn’t easy to find.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;Reason 3: Your Brand Is Weaker on What the User Asked for&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT agrees your brand belongs in the conversation, but something it read says your brand falls short on the exact thing the user asked for. &lt;/span&gt;&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;Most of the time, ChatGPT keeps a brand like this and adds a caveat. The brand gets cut when its weakness is exactly what the user asked to avoid.&amp;nbsp;&lt;/p&gt; 
&lt;p style="font-weight: bold;"&gt;Effort to fix:&amp;nbsp;&lt;span style="font-weight: normal;"&gt;High. This one is harder to fix, because it’s about the brand, not just the page.&lt;/span&gt;&lt;span style="font-weight: normal;"&gt; Look at the complaints ChatGPT repeats about your brand and ask yourself "are these complaints accurate?" and if so, "are they something we want to fix?". &lt;/span&gt;&lt;span style="font-weight: normal;"&gt;For complaints you want to take on, figure out what actually needs to happen and then back it up with product updates, test results, reviews, and third-party coverage.&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;p style="font-weight: bold;"&gt;&lt;span style="font-weight: normal;"&gt;In this AI search era, we've noted brands have to actually "be about it" and not just talk about it, meaning a single page ranking in Google won't fix an issue. &lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;Reason 4: ChatGPT Doesn’t Fully Trust Your Brand&lt;/h3&gt; 
&lt;p&gt;Even when your site says the right things, ChatGPT looks for other sources to back them up: reviews, testing sites, industry lists, and analysts. If it can’t find that, or it finds reasons to doubt you, it may leave you out.&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT often tells itself which sources count. In the Georgia insurance example, it wrote: "Using authoritative sources is key here, so I'll search the official Georgia insurance page," and in another example, it went further: "SEO-driven listicles and self-published content dominate the search results, so I need to assess official agency sites and public evidence, maybe even customer reviews."&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;ChatGPT is telling itself to look past the programmatic, self-published content&lt;/span&gt; that has been built for AI citations over the past year or two. We’ve seen hints of this before: ChatGPT’s listicle citations&amp;nbsp;&lt;/span&gt;&lt;a href="https://www.seerinteractive.com/insights/the-listicle-window-is-closing-in-ai-search-30-decline-mom"&gt;fell 30% from December 2025 to January 2026&lt;/a&gt;&lt;span&gt; in our earlier study.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT rarely says where a trust concern comes from. When it does, it’s usually about how a brand presents itself (self-ranking, self-promotion, unclear pricing) rather than its reviews.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;A self-promoting brand can stay in but lose its recommendation: "[An SEO and digital PR agency] offers a blend of technical and PR services but is self-promotional, so I'll clarify its position without ranking it."&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;span style="font-weight: bold;"&gt;Effort to fix: &lt;/span&gt;&lt;span style="font-weight: normal;"&gt;Medium. Avoid low-quality, programmatic practices and focus on creating content that is actually valuable for humans. We’ve also &lt;a href="https://www.seerinteractive.com/insights/study-of-800k-ai-responses-how-reviews-shape-brand-presence-in-ai-search"&gt;&lt;u&gt;&lt;span&gt;seen review profiles correlate with significantly higher AI citation&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; rates across all platforms, verticals, and brand sizes. &lt;/span&gt;Build up your review profiles and third-party validation on the sites ChatGPT searches for the topics you care about. &lt;/span&gt;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;Reason 5: You’re at the Bottom of ChatGPT’s List&lt;/h3&gt; 
&lt;p&gt;Listing five brands was ChatGPT’s default in our research. When its notes name a list size, it’s five far more often than any other number, and GPT-6 answers usually land on five or six brands. If you’re sixth in line, you’re on the bubble.&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;Knowing how many brands ChatGPT plans to include tells you how close or far you are from being mentioned. In the Georgia home insurance example, ChatGPT said: “I’m considering a list of five insurance companies.” For a price-tracking prompt: “I think we can list about five options.”&amp;nbsp;&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;When ChatGPT says it’s trimming, that brand almost always disappears. In our analysis, &lt;span style="font-weight: bold;"&gt;9 of the 10 brands it flagged this way didn’t make the final response&lt;/span&gt;.&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Effort to fix:&amp;nbsp;&lt;/span&gt;&lt;/strong&gt;&lt;span style="font-weight: normal;"&gt;Depends on the other four. You can’t make the list longer. The way to move up is to fix the other four reasons.&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;What to Do Next&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Start by running the prompts you actually care about in ChatGPT and reading the reasoning. You’ll get a much clearer idea of why your brand is or isn’t included. Keep in mind that ChatGPT’s interface ranges in how much reasoning is shown, which is why we use the API to get this information at scale.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you haven’t used the API before, test out the &lt;/span&gt;&lt;a href="https://platform.openai.com/chat/edit?models=gpt-6-luna"&gt;&lt;u&gt;&lt;span&gt;OpenAI Playground&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; first. Click into the Assistant’s ‘thought’ output and you’ll see the reasoning summary and rules ChatGPT is setting for itself. You can also click into the search outputs to find fan-out queries and sources used.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;As you’re checking each prompt, look for three things:&lt;/span&gt;&lt;/p&gt; 
&lt;ol&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;The sources it says it will trust. Are you on them?&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;The must-haves it names. Does your site state them plainly?&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;How many brands it plans to list, and whether you’re named and then dropped.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ol&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;The quick win&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;: Write down every must-have ChatGPT names (licenses, certifications, integrations, local availability). If you have them, make sure they’re stated plainly on your site.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Once you add those must-haves, monitor the AI performance of those pages and your brand’s visibility and recommendation rate in AI responses. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The reasoning won’t explain every response, and it won’t always be 1:1 with what users see in the app. But it gives you a much clearer view into the criteria ChatGPT is using to decide which brands make the list, and that’s something you can act on.&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Want help understanding why ChatGPT isn’t naming your brand? Seer can help you pinpoint why you’re being excluded and work to get you on ChatGPT’s shortlist. &lt;a href="https://www.seerinteractive.com/contact"&gt;&lt;u&gt;Contact us&lt;/u&gt;&lt;/a&gt; to chat about it.&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-reasoning" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/chatgpt-reasoning.png" alt="Why Doesn’t Your Brand Appear in ChatGPT Responses? Check Its Reasoning" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;We ran 20 prompts through OpenAI's API and tracked every step and reasoning summary to understand why it picked some brands and ignored others&amp;nbsp;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;In our prompt runs, fewer than 1 in 10 brands were named before ChatGPT runs a single search (AKA around step 4 in the process)&lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;ChatGPT searches brand websites infrequently after its first search round (13% of queries) but most of the time in later rounds (75%)&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;We identified 5 main reasons a brand gets excluded: prompt-brand mismatch, brand website is missing info, known brand weakness, trust concerns, or too far down the shortlist&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt;   
&lt;h2&gt;&lt;span&gt;How Does ChatGPT Decide Which Brands Make the Cut?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Wouldn’t it be nice if we knew where ChatGPT citations came from? Or why ChatGPT searches for those specific fanout queries? For over three years now, we’ve been peeling back the layers to better understand how ChatGPT responses come together. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;First, we tracked AI visibility to understand whether your brand shows up at all. Then we focused on citation data and which brands were being sourced. After that we moved on to query fan-outs, which show how those citations are being selected.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;But many of the discussions we’ve seen talk about fan-outs as if those searches happen simultaneously. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;When you dig into ChatGPT’s reasoning, &lt;span style="font-weight: bold;"&gt;you can see a chain of thought that runs from the prompt all the way to the final response&lt;/span&gt;. Looking at the steps of this reasoning chain can help you understand which brands get mentioned, and why. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In this post, we’ll walk through how ChatGPT’s reasoning works, reasons that explain why your brand isn’t showing up, and how to fix it.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;ChatGPT’s Reasoning Is a Signal&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Reasoning is exactly what it sounds like. It’s how ChatGPT decides on the criteria for its answer: &lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;What matters for this question&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Which sources to trust&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Which brands should and shouldn’t make the list&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT’s reasoning isn’t that different from how people think. If you were buying an e-bike to commute from North Park to downtown San Diego and ride trails on the weekend, you wouldn’t start by browsing bikes. You’d think about the hills on your route, check whether e-bikes are even allowed on the trails you ride, and only then decide which features matter.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you’re a brand that sells e-bikes, ChatGPT’s reasoning can tell you a lot more than reviewing cited URLs.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT decided the shortlist of options it will mention should come down to a set of criteria, like the terrain of the commute and California’s trail rules. Understanding the reasoning helps you see what those criteria are, so you have a higher likelihood of being visible in those searches and reinforcing signals ChatGPT is searching for on your own site.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In the app, the reasoning is tucked behind “Worked for 34s” and shown in short form. That’s why we’re going to use the OpenAI API, which shows a summary of every reasoning step in the chain.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The API is directional to what actually happens in ChatGPT’s user interface. And while these results might not be 1:1 with what users are experiencing, &lt;/span&gt;&lt;strong&gt;&lt;span&gt;why not take action on the direct hints ChatGPT is providing?&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Want to see this for your&amp;nbsp;brand? &lt;/span&gt;&lt;/strong&gt;&lt;span style="font-weight: normal;"&gt;We've created a tool to capture every trace in this study. All you have to do is paste a prompt, add your OpenAI key, and watch ChatGPT reason, search, and trim in real time.&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;The Flow of ChatGPT’s Reasoning &lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;span&gt;To better understand ChatGPT’s reasoning, we ran 20 real prompts on November 4, 2026 through OpenAI’s API and read every reasoning summary ChatGPT wrote along the way. We used GPT-6 Luna at medium reasoning to align with ChatGPT’s default model. Medium reasoning was used due to its response and citation results corresponding to the app at the closest rate to other model settings tested.&lt;/span&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;span&gt;What we found is that &lt;/span&gt;&lt;strong&gt;&lt;span&gt;ChatGPT runs through several rounds of searching and reasoning before answering&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;. Let’s break down what each of those steps looks like.&lt;/span&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span style="color: #000000;"&gt;&lt;strong&gt;Step 1: The Prompt&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;This is the start of it all. ChatGPT starts its reasoning after a user submits a prompt.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;One of the prompts we looked at was about home insurance in Georgia:&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span&gt;“Which home insurance companies are most recommended for homeowners in Georgia, and why do they stand out?”&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 2: The First Reasoning Step, Where ChatGPT Builds the Plan and the Source Rules&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Depending on the prompt and your industry, ChatGPT can name those credible sources outright. &lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span style="white-space-collapse: preserve;"&gt;For our Georgia home insurance prompt, it planned to: "check credible sources, including the state insurance regulator, a ratings agency, and an insurance association."&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 3: Broad Searches&lt;/strong&gt;&lt;span style="white-space-collapse: preserve;"&gt; &lt;/span&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Next ChatGPT moves to broad searches, which include a lot of "best of" and "review" queries. These are what surface the listicles commonly discussed in AI citations, along with other articles.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span style="white-space-collapse: preserve;"&gt;For our example, its first round of searches had fan-out queries like “best homeowners insurance companies Georgia 2026 recommendations” and “Georgia homeowners insurance companies customer satisfaction [a ratings agency] 2025.”&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 4: The Second Reasoning Step, Building the Shortlist &lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;In its next reasoning summary, ChatGPT starts naming brands for consideration. In our analysis, it usually names about eight options but ends up mentioning about six in the final response.&lt;/span&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In our prompt runs, &lt;span style="font-weight: bold;"&gt;fewer than 1 in 10 are named before ChatGPT runs a single search&lt;/span&gt;. Most brands (59%) showed up right after the first round of broad searches, which is why retrieval and being visible on retrieved pages matter so much. The rest get added in later rounds.&lt;/span&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;This is also where ChatGPT adds conditions an option has to meet. About half of these reasoning notes add at least one, like “Looking into B2B SaaS routing and CRM integrations is key too.” Then it decides what it still needs to check.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span style="white-space-collapse: preserve;"&gt;After reading a personal-finance site’s Georgia guide, it wrote, “I’m considering a list of five insurance companies,” naming five insurers and flagging a sixth that “might stand out for low complaints.”&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="color: #000000;"&gt;&lt;strong&gt;Step 5: Targeted Searches&lt;/strong&gt;&lt;span style="white-space-collapse: preserve;"&gt; &lt;/span&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Once it has a shortlist, ChatGPT sends out more searches aimed at those specific brands, often using the site: operator to search just one website. In GPT-6’s first round of searching, only 13% of queries do this. In later rounds, 75% do. It finds brands on broad searches, then often goes to their own websites to check them.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span&gt;In our Georgia example, the targeted check went directly to another personal-finance site’s Georgia guide to verify its shortlist.&lt;/span&gt;&lt;/pre&gt; 
&lt;br&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 6: The Last Reasoning Step, Where ChatGPT Trims the List&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Before it writes the response, ChatGPT writes one final note where it trims the list and excludes some of the brands it found. In our prompt set, 38% of these final notes give a reason to drop or exclude an option from the set it will show the user.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span&gt;Right before writing the response, it wrote, “[An insurer] could be optional, but we might trim that,” and decided to say there’s “no universally best insurance” rather than crowning one winner.&lt;/span&gt;&lt;/pre&gt; 
&lt;h3&gt;&lt;span style="white-space-collapse: preserve; color: #000000;"&gt;&lt;strong&gt;Step 7: The Response&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Based on all of its searching and reasoning, ChatGPT serves the final output that users see.&lt;/span&gt;&lt;/p&gt; 
&lt;pre&gt;&lt;span&gt;The final response recommended the five insurers from ChatGPT's shortlist. The sixth, the one it said it “might trim,”&lt;/span&gt;&lt;strong&gt;&lt;span&gt; didn’t make it&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;.&lt;/span&gt;&lt;/pre&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;So, Why Does ChatGPT Exclude a Brand?&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p style="font-weight: normal;"&gt;After reviewing the flow from prompt to response for the prompts analyzed, we identified five primary reasons why ChatGPT excludes a brand.&lt;/p&gt; 
&lt;h3&gt;Reason 1: You Don't Fit the Exact Prompt&lt;/h3&gt; 
&lt;p&gt;Your brand isn’t a bad option, it’s just the wrong kind of option for this prompt. Your brand is built for a different buyer or use case, or you don’t serve the location the user asked about.&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT sets these conditions early. For a Baltimore HVAC prompt it wrote: “I should confirm service area and current ratings, especially for the Baltimore area.” &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt; In one case it even planned to include “two smaller local options in the quote list, even if that means leaving out a higher-ranking company.” &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;You’re not being compared against other brands at this step. You’re being evaluated against the user’s specific needs, location, and criteria. If you don’t meet them, you’re out before the comparison even starts. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Effort to fix: &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;High, often not fixable. Your brand either meets the user’s needs or it doesn’t. If the user wants a local option and you aren’t one, this isn’t a topic to prioritize.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;Reason 2: Your Site Is Missing the Signals ChatGPT Looks for&lt;/h3&gt; 
&lt;p&gt;ChatGPT goes to your official website to check you against its criteria. If the information isn’t there or isn’t easy to find, it moves on.&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In some cases, ChatGPT will name the exact information it's looking for. ChatGPT's reasoning has specifically said it’s crucial to verify credentials, check for relevant certifications, and find evidence your brand matches the criteria.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt; For the Baltimore HVAC prompt, it wrote: “Maryland requires an &lt;/span&gt;&lt;strong&gt;&lt;span&gt;HVACR license&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; for technical work and an &lt;/span&gt;&lt;strong&gt;&lt;span&gt;MHIC license&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; for home improvement contracts, so it’s crucial to verify these credentials for applicable projects.” &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt; If you’re a contractor, that means listing every license, permit, and certification that applies to your work. Some of them ChatGPT may never check for, but it won’t hurt to have them on your site, and you can’t predict every requirement it’ll look for. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;So brands should know which requirements ChatGPT is looking for. Some brands have them, but the information isn’t easy for users or AI to find, so they’re excluded by default.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Effort to fix: &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Low, this is the one to focus on. This could be the most tactical way to improve your AI performance: surface information you already have that just isn’t easy to find.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;Reason 3: Your Brand Is Weaker on What the User Asked for&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT agrees your brand belongs in the conversation, but something it read says your brand falls short on the exact thing the user asked for. &lt;/span&gt;&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;Most of the time, ChatGPT keeps a brand like this and adds a caveat. The brand gets cut when its weakness is exactly what the user asked to avoid.&amp;nbsp;&lt;/p&gt; 
&lt;p style="font-weight: bold;"&gt;Effort to fix:&amp;nbsp;&lt;span style="font-weight: normal;"&gt;High. This one is harder to fix, because it’s about the brand, not just the page.&lt;/span&gt;&lt;span style="font-weight: normal;"&gt; Look at the complaints ChatGPT repeats about your brand and ask yourself "are these complaints accurate?" and if so, "are they something we want to fix?". &lt;/span&gt;&lt;span style="font-weight: normal;"&gt;For complaints you want to take on, figure out what actually needs to happen and then back it up with product updates, test results, reviews, and third-party coverage.&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;p style="font-weight: bold;"&gt;&lt;span style="font-weight: normal;"&gt;In this AI search era, we've noted brands have to actually "be about it" and not just talk about it, meaning a single page ranking in Google won't fix an issue. &lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;Reason 4: ChatGPT Doesn’t Fully Trust Your Brand&lt;/h3&gt; 
&lt;p&gt;Even when your site says the right things, ChatGPT looks for other sources to back them up: reviews, testing sites, industry lists, and analysts. If it can’t find that, or it finds reasons to doubt you, it may leave you out.&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT often tells itself which sources count. In the Georgia insurance example, it wrote: "Using authoritative sources is key here, so I'll search the official Georgia insurance page," and in another example, it went further: "SEO-driven listicles and self-published content dominate the search results, so I need to assess official agency sites and public evidence, maybe even customer reviews."&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;ChatGPT is telling itself to look past the programmatic, self-published content&lt;/span&gt; that has been built for AI citations over the past year or two. We’ve seen hints of this before: ChatGPT’s listicle citations&amp;nbsp;&lt;/span&gt;&lt;a href="https://www.seerinteractive.com/insights/the-listicle-window-is-closing-in-ai-search-30-decline-mom"&gt;fell 30% from December 2025 to January 2026&lt;/a&gt;&lt;span&gt; in our earlier study.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT rarely says where a trust concern comes from. When it does, it’s usually about how a brand presents itself (self-ranking, self-promotion, unclear pricing) rather than its reviews.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;A self-promoting brand can stay in but lose its recommendation: "[An SEO and digital PR agency] offers a blend of technical and PR services but is self-promotional, so I'll clarify its position without ranking it."&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;span style="font-weight: bold;"&gt;Effort to fix: &lt;/span&gt;&lt;span style="font-weight: normal;"&gt;Medium. Avoid low-quality, programmatic practices and focus on creating content that is actually valuable for humans. We’ve also &lt;a href="https://www.seerinteractive.com/insights/study-of-800k-ai-responses-how-reviews-shape-brand-presence-in-ai-search"&gt;&lt;u&gt;&lt;span&gt;seen review profiles correlate with significantly higher AI citation&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; rates across all platforms, verticals, and brand sizes. &lt;/span&gt;Build up your review profiles and third-party validation on the sites ChatGPT searches for the topics you care about. &lt;/span&gt;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;Reason 5: You’re at the Bottom of ChatGPT’s List&lt;/h3&gt; 
&lt;p&gt;Listing five brands was ChatGPT’s default in our research. When its notes name a list size, it’s five far more often than any other number, and GPT-6 answers usually land on five or six brands. If you’re sixth in line, you’re on the bubble.&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;Knowing how many brands ChatGPT plans to include tells you how close or far you are from being mentioned. In the Georgia home insurance example, ChatGPT said: “I’m considering a list of five insurance companies.” For a price-tracking prompt: “I think we can list about five options.”&amp;nbsp;&lt;/p&gt; 
&lt;p style="font-weight: normal;"&gt;When ChatGPT says it’s trimming, that brand almost always disappears. In our analysis, &lt;span style="font-weight: bold;"&gt;9 of the 10 brands it flagged this way didn’t make the final response&lt;/span&gt;.&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Effort to fix:&amp;nbsp;&lt;/span&gt;&lt;/strong&gt;&lt;span style="font-weight: normal;"&gt;Depends on the other four. You can’t make the list longer. The way to move up is to fix the other four reasons.&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;What to Do Next&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Start by running the prompts you actually care about in ChatGPT and reading the reasoning. You’ll get a much clearer idea of why your brand is or isn’t included. Keep in mind that ChatGPT’s interface ranges in how much reasoning is shown, which is why we use the API to get this information at scale.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you haven’t used the API before, test out the &lt;/span&gt;&lt;a href="https://platform.openai.com/chat/edit?models=gpt-6-luna"&gt;&lt;u&gt;&lt;span&gt;OpenAI Playground&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; first. Click into the Assistant’s ‘thought’ output and you’ll see the reasoning summary and rules ChatGPT is setting for itself. You can also click into the search outputs to find fan-out queries and sources used.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;As you’re checking each prompt, look for three things:&lt;/span&gt;&lt;/p&gt; 
&lt;ol&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;The sources it says it will trust. Are you on them?&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;The must-haves it names. Does your site state them plainly?&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;How many brands it plans to list, and whether you’re named and then dropped.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ol&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;The quick win&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;: Write down every must-have ChatGPT names (licenses, certifications, integrations, local availability). If you have them, make sure they’re stated plainly on your site.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Once you add those must-haves, monitor the AI performance of those pages and your brand’s visibility and recommendation rate in AI responses. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The reasoning won’t explain every response, and it won’t always be 1:1 with what users see in the app. But it gives you a much clearer view into the criteria ChatGPT is using to decide which brands make the list, and that’s something you can act on.&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Want help understanding why ChatGPT isn’t naming your brand? Seer can help you pinpoint why you’re being excluded and work to get you on ChatGPT’s shortlist. &lt;a href="https://www.seerinteractive.com/contact"&gt;&lt;u&gt;Contact us&lt;/u&gt;&lt;/a&gt; to chat about it.&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fchatgpt-reasoning&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Author: Nick Haigler</category>
      <category>AI</category>
      <pubDate>Wed, 18 Nov 2026 18:42:00 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/chatgpt-reasoning</guid>
      <dc:date>2026-11-18T18:42:00Z</dc:date>
      <dc:creator>Nick Haigler</dc:creator>
    </item>
    <item>
      <title>ChatGPT Ads UX Study: Findings from Real AI Search Sessions</title>
      <link>https://www.seerinteractive.com/insights/chatgpt-ads-inside-answers-what-161-ai-search-sessions-show</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-ads-inside-answers-what-161-ai-search-sessions-show" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Blog%20Update%20%231%20(10).png" alt="ChatGPT Ads UX Study: Findings from Real AI Search Sessions" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;p&gt;We watched 161 recorded AI search sessions and logged every ad that showed up inside a ChatGPT answer. Most of the people we asked told us they hadn't seen any.&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-ads-inside-answers-what-161-ai-search-sessions-show" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Blog%20Update%20%231%20(10).png" alt="ChatGPT Ads UX Study: Findings from Real AI Search Sessions" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;p&gt;We watched 161 recorded AI search sessions and logged every ad that showed up inside a ChatGPT answer. Most of the people we asked told us they hadn't seen any.&lt;/p&gt;   
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fchatgpt-ads-inside-answers-what-161-ai-search-sessions-show&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Creative Services</category>
      <category>Author: Denise Baginski</category>
      <category>Consumer Insights</category>
      <category>Author: Andrea Haley</category>
      <category>AI</category>
      <category>chatgpt</category>
      <category>AI Study</category>
      <pubDate>Thu, 12 Nov 2026 19:00:00 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/chatgpt-ads-inside-answers-what-161-ai-search-sessions-show</guid>
      <dc:date>2026-11-12T19:00:00Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>2025 - ChatGPT named 21 competitors &amp; us 0 times 2026 - Them 0, us 65%</title>
      <link>https://www.seerinteractive.com/insights/chatgpt-named-21-competitors-in-2025-and-us-0-today-many-are-at-0-we-are-at-65-percent</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-named-21-competitors-in-2025-and-us-0-today-many-are-at-0-we-are-at-65-percent" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/image-png-Nov-12-2026-05-18-44-9604-PM.png" alt="2025 - ChatGPT named 21 competitors &amp;amp; us 0 times 2026 - Them 0, us 65%" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2&gt;In 2025 ChatGPT rewarded our competitors for ranking themselves in listicles.&lt;/h2&gt; 
&lt;h3&gt;We bet on humans instead, this is the data on what happened.&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;In February 2025 I recorded this video:&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-named-21-competitors-in-2025-and-us-0-today-many-are-at-0-we-are-at-65-percent" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/image-png-Nov-12-2026-05-18-44-9604-PM.png" alt="2025 - ChatGPT named 21 competitors &amp;amp; us 0 times 2026 - Them 0, us 65%" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2&gt;In 2025 ChatGPT rewarded our competitors for ranking themselves in listicles.&lt;/h2&gt; 
&lt;h3&gt;We bet on humans instead, this is the data on what happened.&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;In February 2025 I recorded this video:&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fchatgpt-named-21-competitors-in-2025-and-us-0-today-many-are-at-0-we-are-at-65-percent&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>author-wil-reynolds</category>
      <category>GEO</category>
      <pubDate>Thu, 12 Nov 2026 18:01:18 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/chatgpt-named-21-competitors-in-2025-and-us-0-today-many-are-at-0-we-are-at-65-percent</guid>
      <dc:date>2026-11-12T18:01:18Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>Testing llms.txt vs. AI Info Pages: GEO Quick Wins or Waste of Time?</title>
      <link>https://www.seerinteractive.com/insights/llms.txt-vs-ai-info-page</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/llms.txt-vs-ai-info-page" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/llms.txt%20vs%20AI%20info%20page%20blog-1.png" alt="Testing llms.txt vs. AI Info Pages: GEO Quick Wins or Waste of Time?" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;I compared llms.txt and AI info pages to determine if these were easy GEO wins or tactics without data proof&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;llms.txt files failed to perform across several parameters: 0 for 2 on formal client tests, and almost no citation presence&lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;AI info pages performed for 3 of 4 clients, and Seer's AI info page continues to be consistently cited a year after going live&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;A page bots never fetch can’t change an answer: one client’s AI info page got no traction as Markdown or as HTML, because AI agents weren’t requesting it&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;In terms of which is more likely to be cited in AI responses, the AI info page is your winner&lt;/li&gt; 
&lt;/ul&gt;  
&lt;h2&gt;&lt;span&gt;Are llms.txt and AI Info Pages Worth Your Time?&amp;nbsp;&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;When the GEO race started, everyone was guessing about what worked - even us data wranglers. Without evidence of what worked &lt;/span&gt;&lt;em&gt;&lt;span&gt;well&lt;/span&gt;&lt;/em&gt;&lt;span&gt;, we settled for what worked &lt;/span&gt;&lt;em&gt;&lt;span&gt;quickly&lt;/span&gt;&lt;/em&gt;&lt;span&gt;. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Two of those tactics keep landing in my inbox: llms.txt and the AI info page. They’re often lumped together because they look alike and are "for the bots." Both are also pitched as low-lift and quick wins. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;But llms.txt and the AI info page do different jobs, and in our testing they had very different levels of success: &lt;/span&gt;&lt;strong&gt;&lt;span&gt;the AI info page got picked up for 3 of 4 clients, llms.txt for none.&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here's how they differ, what happened when we tested them, and when I'd reach for each.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;llms.txt vs. AI Info Page: What’s the Difference?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;llms.txt and AI info pages both get pitched as ways for AI answer engines to find and fact-check your brand. But they serve different purposes, and bots may or may not read them.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here are our own examples from seerinteractive.com:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;a href="https://www.seerinteractive.com/llms.txt"&gt;&lt;strong&gt;&lt;u&gt;&lt;span&gt;Seer’s llms.txt page&lt;/span&gt;&lt;/u&gt;&lt;/strong&gt;&lt;/a&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;a href="https://www.seerinteractive.com/ai-information"&gt;&lt;strong&gt;&lt;u&gt;&lt;span&gt;Seer's AI Info Page&lt;/span&gt;&lt;/u&gt;&lt;/strong&gt;&lt;/a&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;llms.txt started as a &lt;/span&gt;&lt;a href="https://llmstxt.org/"&gt;&lt;u&gt;&lt;span&gt;community proposal&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; for a plain-text file that points models to a site’s most useful pages. The AI info page has no spec at all. It’s just a web page written for a different reader.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;And here’s a quick overview of how llms.txt and the AI info page differ:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Bots browse your site differently from how people do: they search an index, or fetch a specific URL live when a question calls for it. An AI info page fits that path, while an llms.txt file gets largely ignored unless a crawler goes looking for it. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;When we tested the two tactics, the AI info page worked for Seer and 3 of the 4 clients who've tested it. &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Conversely, llms.txt failed to draw meaningful bot traffic, let alone steer it.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;In Our Tests, AI Info Pages Outperformed llms.txt&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;We have a GEO Test &amp;amp; Learn Library that tracks every completed GEO test we've run for our clients - 70 completed GEO tests across 17 clients as of October 5, 2026. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Our findings from client tests of llms.txt and AI info pages were as follows:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;0 for 2 on tests targeting llms.txt&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;3 for 4 on AI info page tests&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Counting a head-to-head test where both launched the same day, llms.txt is 0 for 3&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;When we ran the tests for ourselves at Seer, we saw the same effects.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;llms.txt Saw Almost No Traction Among Bots&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;In 2025 we shipped llms.txt for several clients. It was cheap and LinkedIn was loud about it, so we decided to test it. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Overall, our findings on llms.txt were underwhelming:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Two clients ran formal tests, and neither logged a single ChatGPT request for the file &lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Across nine total clients, answer agents barely asked for it and the net impact on AI citations was negligible&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;We even tried a different perspective, analyzing the performance of llms.txt in AI answers. &lt;/span&gt;&lt;strong&gt;&lt;span&gt;Of about 9.2M AI answers across 100 client data sets, only 73 cited an llms.txt URL&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;, mostly developer docs. Of those, 15 were Gemini inventing yourdomain.com/llms.txt while giving GEO advice.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;We’re not the only ones seeing it. &lt;/span&gt;&lt;a href="https://ahrefs.com/blog/llmstxt-study/"&gt;Ahrefs&lt;/a&gt;&lt;span&gt; found that of roughly 38,000 sites with a valid file, 97% got no requests for it in May 2026. An &lt;/span&gt;&lt;a href="https://www.searchenginejournal.com/llms-txt-shows-no-clear-effect-on-ai-citations-based-on-300k-domains/561542/"&gt;&lt;u&gt;&lt;span&gt;SE Ranking study of about 300,000 domains&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; found no clear effect on AI citations. And &lt;/span&gt;&lt;a href="https://developers.google.com/search/docs/fundamentals/ai-optimization-guide"&gt;Google’s own guidance&lt;/a&gt;&lt;span&gt; says Search ignores llms.txt.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Once the results came in, we stopped recommending llms.txt as a GEO tactic. Performance didn't move, LLMs didn’t cite the files, and the logs showed AI crawlers weren't requesting them. &lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;AI Info Pages Performed for Us and Our Clients&lt;/span&gt;&lt;/h3&gt; 
&lt;p style="font-weight: normal;"&gt;Often we say that good science is both repeatable and reproducible. So we ran AI info page tests just like the llms.txt tests, and found a very different picture.&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;First, the bots actually go get the page. Here’s how agent requests for the AI info page compared with llms.txt on the two client sites where we can see bot logs, and on our own. robots.txt is in the table as a yardstick, since nearly every crawler checks it:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Every site in our logs with an AI info page saw far more agent requests for the page than for llms.txt, including the developer platform that runs both.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Second, what’s on the page shows up in the answers. &lt;/span&gt;&lt;strong&gt;&lt;span&gt;Four clients have tested an AI info page, and AI answers picked it up in three&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Energy technology company:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; 21 of 57 tracked facts (37%) showed up in ChatGPT answers about a month after launch&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Enterprise data-protection software company:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Launched an AI info page and an llms.txt the same day; Gemini and Perplexity cited the page and never the file, though that’s an early read on about 10 answers in three weeks&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Accounts receivable software company:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The client’s team called it a win&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Communications SaaS company:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Bots never fetched the page, so it never had a chance to change an answer (more on that below)&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;Audience and industry framing stuck best, and proprietary product names stuck least. That's a good reminder to use the words your buyers do.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Outside of our Library, we also tracked a financial data and ratings firm. The page drew 439 requests from ChatGPT-User, 74% of all requests to it. ChatGPT-User fires when a real person’s question sends ChatGPT to read a page live. In a 10-answer sample of branded ChatGPT responses, about half picked up some of the page’s language, most of it partial matches.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;No Fetch, No Result&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;The miss was the most useful result of the bunch. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The communications SaaS company’s Markdown (.md) version of the AI info page drew zero AI crawler requests in its final week, and only 1 of 30 tracked facts showed up in answers. The team relaunched it as HTML in January 2026 and got 0 of 30 on the same prompts. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;Changing the format didn’t fix it, because bots weren’t fetching the page. A page that never gets fetched can’t change an answer, regardless of what's on it.&lt;/p&gt; 
&lt;p&gt;We still recommend HTML, because every AI info page that worked in our data was HTML, and HTML is what answer engines search for and fetch live. But this client shows that the first thing to check is the logs, not the content.&lt;/p&gt; 
&lt;h3&gt;A Year In, Seer’s AI Info Page Still Owns the Who-Are-You Answer&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;We looked at a year of data to see whether our AI info page has continued to perform. Seer's AI info page went live in October 2025 and Wil Reynolds has tracked it in Scrunch since mid-November.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;On the most direct brand question there is, "Tell me about Seer Interactive," the AI info page has been cited in about half of answers all year&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;: 50.4% on average from late November through March and 52.5% from June through September, across 403 answers. It held steady even when citations on our other brand prompts dropped in late January.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Then we split the branded prompts in two. &lt;strong&gt;Plain brand prompts&lt;/strong&gt; ask about Seer by name and nothing else, like “Tell me about Seer Interactive.” &lt;strong&gt;Persona prompts&lt;/strong&gt; still name Seer, but they come in a buyer’s voice, with a role and a problem attached: “As a CMO whose paid media costs keep rising while organic results decline, I want to know how Seer Interactive builds integrated strategies that reduce paid dependency.” &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The first is a who-are-you question. The second is a should-I-hire-you question, and the model pulls from a lot more of the web to answer it.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Plain brand prompts:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; On the six we track, the citation rate fell from 57% to 37% in a single step on January 26, then slid into the high 20s by late March.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Persona prompts:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; On the 41 persona-style prompts we switched to in April, citation rate fell from 19.5% to 7.8% around the start of June and has sat between roughly 5% and 13% since.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;So the page holds the who-are-you answer and loses ground on should-I-hire-you. That’s the line between what an AI info page does and doesn’t do. It isn’t set and forget, either: the core answer held while the broader branded share drifted down.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Now the Bayesian caveat. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The llms.txt results are the better-controlled ones: subfolder comparisons and server logs, though each has only two to three weeks of post-launch data. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;My confidence there is high. llms.txt doesn't get fetched. The AI info page wins are thinner. None had a control, most were read on a single platform, and one is an early read on about 10 answers. I'm moderately confident in the direction, and I expect to update. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Our own backlog has already voted, with 21 AI info page tests proposed or in flight and exactly one llms.txt test.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;llms.txt or AI Info Page: Which Should You Use?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Stop asking "which tactic should I use?" and start asking "what problem am I trying to fix?" For most brands I talk to, the problem is that AI gets their brand information or narrative wrong. That calls for an AI info page.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here’s a breakdown of what to focus on based on what you’re trying to accomplish:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;What if a client wants llms.txt "just in case"? Fine. It's about an hour of work, and it won't hurt anything. Ship it, link your AI info page from it, and move on. Just don't count it as your GEO strategy.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;What I'd Do Today&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;If you only do one thing based on this blog, build an AI info page. The condensed playbook:&lt;/span&gt;&lt;/p&gt; 
&lt;ol&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Find the gaps.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Run your branded prompts and write down what AI gets wrong or leaves out.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Write plain facts, with corrections spelled out.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Short sections, dated numbers, and your buyers' words instead of your internal product names.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Publish it as HTML.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; A clean URL, a 200 status, no noindex, and no JavaScript needed to see the text. Link it from the footer and add it to your sitemap.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Check the logs before you judge anything.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; No fetches from ChatGPT-User, PerplexityBot, or similar agents means no test.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Measure language adoption on branded prompts.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; First look at 4 weeks, make the call at 12, then refresh it every quarter.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ol&gt; 
&lt;p&gt;&lt;span&gt;All of this fits into a bigger shift in how the web works these days. As we move fact-finding and research to AI tools and Answer engines, the focus goes from how to organize information for human consumption (clean UX, semantic markup, responsive design for every conceivable screen size or device type) to AI consumption (schema, .md files, etc.).&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;My colleagues, to their credit, have been saying so already. Alisa Scharf argues you should &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/insights/stop-chasing-ai-rankings-before-you-fix-how-llms-see-your-brand"&gt;&lt;u&gt;&lt;span&gt;fix how LLMs see your brand before chasing AI rankings&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt;, and John Lovett lays out how to write those facts down in a &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/insights/brand-canon"&gt;&lt;u&gt;&lt;span&gt;brand canon&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt;.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you want help figuring out where to focus your GEO efforts, Seer can build and implement a strategy for you. We can help create your brand's AI info page or llms.txt, or even work with you to build a brand canon. &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/contact"&gt;&lt;u&gt;&lt;span&gt;Chat with us&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; to learn more.&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/llms.txt-vs-ai-info-page" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/llms.txt%20vs%20AI%20info%20page%20blog-1.png" alt="Testing llms.txt vs. AI Info Pages: GEO Quick Wins or Waste of Time?" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;I compared llms.txt and AI info pages to determine if these were easy GEO wins or tactics without data proof&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;llms.txt files failed to perform across several parameters: 0 for 2 on formal client tests, and almost no citation presence&lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;AI info pages performed for 3 of 4 clients, and Seer's AI info page continues to be consistently cited a year after going live&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;A page bots never fetch can’t change an answer: one client’s AI info page got no traction as Markdown or as HTML, because AI agents weren’t requesting it&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;In terms of which is more likely to be cited in AI responses, the AI info page is your winner&lt;/li&gt; 
&lt;/ul&gt;  
&lt;h2&gt;&lt;span&gt;Are llms.txt and AI Info Pages Worth Your Time?&amp;nbsp;&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;When the GEO race started, everyone was guessing about what worked - even us data wranglers. Without evidence of what worked &lt;/span&gt;&lt;em&gt;&lt;span&gt;well&lt;/span&gt;&lt;/em&gt;&lt;span&gt;, we settled for what worked &lt;/span&gt;&lt;em&gt;&lt;span&gt;quickly&lt;/span&gt;&lt;/em&gt;&lt;span&gt;. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Two of those tactics keep landing in my inbox: llms.txt and the AI info page. They’re often lumped together because they look alike and are "for the bots." Both are also pitched as low-lift and quick wins. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;But llms.txt and the AI info page do different jobs, and in our testing they had very different levels of success: &lt;/span&gt;&lt;strong&gt;&lt;span&gt;the AI info page got picked up for 3 of 4 clients, llms.txt for none.&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here's how they differ, what happened when we tested them, and when I'd reach for each.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;llms.txt vs. AI Info Page: What’s the Difference?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;llms.txt and AI info pages both get pitched as ways for AI answer engines to find and fact-check your brand. But they serve different purposes, and bots may or may not read them.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here are our own examples from seerinteractive.com:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;a href="https://www.seerinteractive.com/llms.txt"&gt;&lt;strong&gt;&lt;u&gt;&lt;span&gt;Seer’s llms.txt page&lt;/span&gt;&lt;/u&gt;&lt;/strong&gt;&lt;/a&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;a href="https://www.seerinteractive.com/ai-information"&gt;&lt;strong&gt;&lt;u&gt;&lt;span&gt;Seer's AI Info Page&lt;/span&gt;&lt;/u&gt;&lt;/strong&gt;&lt;/a&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;llms.txt started as a &lt;/span&gt;&lt;a href="https://llmstxt.org/"&gt;&lt;u&gt;&lt;span&gt;community proposal&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; for a plain-text file that points models to a site’s most useful pages. The AI info page has no spec at all. It’s just a web page written for a different reader.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;And here’s a quick overview of how llms.txt and the AI info page differ:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Bots browse your site differently from how people do: they search an index, or fetch a specific URL live when a question calls for it. An AI info page fits that path, while an llms.txt file gets largely ignored unless a crawler goes looking for it. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;When we tested the two tactics, the AI info page worked for Seer and 3 of the 4 clients who've tested it. &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Conversely, llms.txt failed to draw meaningful bot traffic, let alone steer it.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;In Our Tests, AI Info Pages Outperformed llms.txt&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;We have a GEO Test &amp;amp; Learn Library that tracks every completed GEO test we've run for our clients - 70 completed GEO tests across 17 clients as of October 5, 2026. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Our findings from client tests of llms.txt and AI info pages were as follows:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;0 for 2 on tests targeting llms.txt&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;3 for 4 on AI info page tests&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Counting a head-to-head test where both launched the same day, llms.txt is 0 for 3&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;When we ran the tests for ourselves at Seer, we saw the same effects.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;llms.txt Saw Almost No Traction Among Bots&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;In 2025 we shipped llms.txt for several clients. It was cheap and LinkedIn was loud about it, so we decided to test it. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Overall, our findings on llms.txt were underwhelming:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Two clients ran formal tests, and neither logged a single ChatGPT request for the file &lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Across nine total clients, answer agents barely asked for it and the net impact on AI citations was negligible&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;We even tried a different perspective, analyzing the performance of llms.txt in AI answers. &lt;/span&gt;&lt;strong&gt;&lt;span&gt;Of about 9.2M AI answers across 100 client data sets, only 73 cited an llms.txt URL&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;, mostly developer docs. Of those, 15 were Gemini inventing yourdomain.com/llms.txt while giving GEO advice.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;We’re not the only ones seeing it. &lt;/span&gt;&lt;a href="https://ahrefs.com/blog/llmstxt-study/"&gt;Ahrefs&lt;/a&gt;&lt;span&gt; found that of roughly 38,000 sites with a valid file, 97% got no requests for it in May 2026. An &lt;/span&gt;&lt;a href="https://www.searchenginejournal.com/llms-txt-shows-no-clear-effect-on-ai-citations-based-on-300k-domains/561542/"&gt;&lt;u&gt;&lt;span&gt;SE Ranking study of about 300,000 domains&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; found no clear effect on AI citations. And &lt;/span&gt;&lt;a href="https://developers.google.com/search/docs/fundamentals/ai-optimization-guide"&gt;Google’s own guidance&lt;/a&gt;&lt;span&gt; says Search ignores llms.txt.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Once the results came in, we stopped recommending llms.txt as a GEO tactic. Performance didn't move, LLMs didn’t cite the files, and the logs showed AI crawlers weren't requesting them. &lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;AI Info Pages Performed for Us and Our Clients&lt;/span&gt;&lt;/h3&gt; 
&lt;p style="font-weight: normal;"&gt;Often we say that good science is both repeatable and reproducible. So we ran AI info page tests just like the llms.txt tests, and found a very different picture.&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;First, the bots actually go get the page. Here’s how agent requests for the AI info page compared with llms.txt on the two client sites where we can see bot logs, and on our own. robots.txt is in the table as a yardstick, since nearly every crawler checks it:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Every site in our logs with an AI info page saw far more agent requests for the page than for llms.txt, including the developer platform that runs both.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Second, what’s on the page shows up in the answers. &lt;/span&gt;&lt;strong&gt;&lt;span&gt;Four clients have tested an AI info page, and AI answers picked it up in three&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Energy technology company:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; 21 of 57 tracked facts (37%) showed up in ChatGPT answers about a month after launch&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Enterprise data-protection software company:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Launched an AI info page and an llms.txt the same day; Gemini and Perplexity cited the page and never the file, though that’s an early read on about 10 answers in three weeks&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Accounts receivable software company:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The client’s team called it a win&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Communications SaaS company:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Bots never fetched the page, so it never had a chance to change an answer (more on that below)&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;Audience and industry framing stuck best, and proprietary product names stuck least. That's a good reminder to use the words your buyers do.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Outside of our Library, we also tracked a financial data and ratings firm. The page drew 439 requests from ChatGPT-User, 74% of all requests to it. ChatGPT-User fires when a real person’s question sends ChatGPT to read a page live. In a 10-answer sample of branded ChatGPT responses, about half picked up some of the page’s language, most of it partial matches.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;No Fetch, No Result&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;The miss was the most useful result of the bunch. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The communications SaaS company’s Markdown (.md) version of the AI info page drew zero AI crawler requests in its final week, and only 1 of 30 tracked facts showed up in answers. The team relaunched it as HTML in January 2026 and got 0 of 30 on the same prompts. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;Changing the format didn’t fix it, because bots weren’t fetching the page. A page that never gets fetched can’t change an answer, regardless of what's on it.&lt;/p&gt; 
&lt;p&gt;We still recommend HTML, because every AI info page that worked in our data was HTML, and HTML is what answer engines search for and fetch live. But this client shows that the first thing to check is the logs, not the content.&lt;/p&gt; 
&lt;h3&gt;A Year In, Seer’s AI Info Page Still Owns the Who-Are-You Answer&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;We looked at a year of data to see whether our AI info page has continued to perform. Seer's AI info page went live in October 2025 and Wil Reynolds has tracked it in Scrunch since mid-November.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;On the most direct brand question there is, "Tell me about Seer Interactive," the AI info page has been cited in about half of answers all year&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;: 50.4% on average from late November through March and 52.5% from June through September, across 403 answers. It held steady even when citations on our other brand prompts dropped in late January.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Then we split the branded prompts in two. &lt;strong&gt;Plain brand prompts&lt;/strong&gt; ask about Seer by name and nothing else, like “Tell me about Seer Interactive.” &lt;strong&gt;Persona prompts&lt;/strong&gt; still name Seer, but they come in a buyer’s voice, with a role and a problem attached: “As a CMO whose paid media costs keep rising while organic results decline, I want to know how Seer Interactive builds integrated strategies that reduce paid dependency.” &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The first is a who-are-you question. The second is a should-I-hire-you question, and the model pulls from a lot more of the web to answer it.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Plain brand prompts:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; On the six we track, the citation rate fell from 57% to 37% in a single step on January 26, then slid into the high 20s by late March.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Persona prompts:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; On the 41 persona-style prompts we switched to in April, citation rate fell from 19.5% to 7.8% around the start of June and has sat between roughly 5% and 13% since.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;So the page holds the who-are-you answer and loses ground on should-I-hire-you. That’s the line between what an AI info page does and doesn’t do. It isn’t set and forget, either: the core answer held while the broader branded share drifted down.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Now the Bayesian caveat. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The llms.txt results are the better-controlled ones: subfolder comparisons and server logs, though each has only two to three weeks of post-launch data. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;My confidence there is high. llms.txt doesn't get fetched. The AI info page wins are thinner. None had a control, most were read on a single platform, and one is an early read on about 10 answers. I'm moderately confident in the direction, and I expect to update. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Our own backlog has already voted, with 21 AI info page tests proposed or in flight and exactly one llms.txt test.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;llms.txt or AI Info Page: Which Should You Use?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Stop asking "which tactic should I use?" and start asking "what problem am I trying to fix?" For most brands I talk to, the problem is that AI gets their brand information or narrative wrong. That calls for an AI info page.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here’s a breakdown of what to focus on based on what you’re trying to accomplish:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;What if a client wants llms.txt "just in case"? Fine. It's about an hour of work, and it won't hurt anything. Ship it, link your AI info page from it, and move on. Just don't count it as your GEO strategy.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;What I'd Do Today&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;If you only do one thing based on this blog, build an AI info page. The condensed playbook:&lt;/span&gt;&lt;/p&gt; 
&lt;ol&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Find the gaps.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Run your branded prompts and write down what AI gets wrong or leaves out.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Write plain facts, with corrections spelled out.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Short sections, dated numbers, and your buyers' words instead of your internal product names.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Publish it as HTML.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; A clean URL, a 200 status, no noindex, and no JavaScript needed to see the text. Link it from the footer and add it to your sitemap.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Check the logs before you judge anything.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; No fetches from ChatGPT-User, PerplexityBot, or similar agents means no test.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Measure language adoption on branded prompts.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; First look at 4 weeks, make the call at 12, then refresh it every quarter.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ol&gt; 
&lt;p&gt;&lt;span&gt;All of this fits into a bigger shift in how the web works these days. As we move fact-finding and research to AI tools and Answer engines, the focus goes from how to organize information for human consumption (clean UX, semantic markup, responsive design for every conceivable screen size or device type) to AI consumption (schema, .md files, etc.).&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;My colleagues, to their credit, have been saying so already. Alisa Scharf argues you should &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/insights/stop-chasing-ai-rankings-before-you-fix-how-llms-see-your-brand"&gt;&lt;u&gt;&lt;span&gt;fix how LLMs see your brand before chasing AI rankings&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt;, and John Lovett lays out how to write those facts down in a &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/insights/brand-canon"&gt;&lt;u&gt;&lt;span&gt;brand canon&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt;.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you want help figuring out where to focus your GEO efforts, Seer can build and implement a strategy for you. We can help create your brand's AI info page or llms.txt, or even work with you to build a brand canon. &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/contact"&gt;&lt;u&gt;&lt;span&gt;Chat with us&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; to learn more.&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fllms.txt-vs-ai-info-page&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Analytics</category>
      <category>AI</category>
      <category>Author: Corey Abramson</category>
      <pubDate>Tue, 10 Nov 2026 17:53:22 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/llms.txt-vs-ai-info-page</guid>
      <dc:date>2026-11-10T17:53:22Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>I Built a Tool to Watch ChatGPT Search, and It Predicted My Next Question</title>
      <link>https://www.seerinteractive.com/insights/chatgpt-search-tool</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-search-tool" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/ChatGPT-search-tool-Hoa-blog.png" alt="I Built a Tool to Watch ChatGPT Search, and It Predicted My Next Question" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;I wanted to understand how ChatGPT searched, so I wrote a script to watch it and tracked 3 random shopping conversations: a carry-on suitcase, a cloud backup tool, and a Spanish app&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;For each topic, I asked an opening question plus two follow-ups in the same chat and saved the data as ChatGPT was building the answer&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;In 2 of the 3 conversations, ChatGPT's first search added a concern I hadn't raised yet but planned to ask about&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;ChatGPT started naming brands during its first search and cited brand sites in that first response for 2 of the 3 conversations&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;In the luggage conversation, ChatGPT cited one review site for 5 of the 9 sources on the first question, and the recommendation lined up with that site's ranking - then ChatGPT looked for further evidence to support that pick in subsequent answers&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;We can use this tool to help clients understand if their brand shows up in the first answer, and what ChatGPT goes looking for when someone pushes back on that answer&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt;  
&lt;h2&gt;&lt;span&gt;De-Mystifying the Black Box of AI Search&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;AI search is a black box. We still don’t know a lot about how it searches, where it pulls information from, or why it cites some sources but not others. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;After &lt;/span&gt;&lt;a href="https://suganthan.com/blog/chatgpt-new-search-language/"&gt;&lt;u&gt;&lt;span&gt;Suganthan Mohanadasan's blog on ChatGPT search language&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; made the rounds in our internal Slack, I was inspired to figure out how to reverse-engineer ChatGPT’s AI search. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I wanted a diagnostic engine that would simulate realistic user conversations and capture the complete retrieval layer, so we could monitor the AI prompts and conversations our clients care about. So I built one.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Instead of just knowing whether a brand shows up in a response, I can see exactly what ChatGPT searches for.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Our team can use this information to see which sites ChatGPT queries to find information about a client’s vertical, analyze why a competitor was cited instead of a client, and even run tactical tests on client content.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Let’s dive into what I built and my initial findings. &lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;I Built A Script to Watch ChatGPT Search in Real Time&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Suganathan’s blog looked at a single query, but I wanted the backend data. Basically, the AI has its own language it speaks when fetching real-time data, and you can see it if you know where to look.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;While ChatGPT builds an answer, the website streams data to your browser, including: &lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;What searches it runs&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;The results that come back &lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Which of those become citations &lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Product records &lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;I'll call that information the stream. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Working with AI coding tools, &lt;/span&gt;&lt;strong&gt;&lt;span&gt;I built a script that types prompts into ChatGPT's website the way a person would&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;. It saves the answer and a screenshot, and it keeps the full stream for every turn.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The other challenge I had to address was Suganthan’s post looked at a single query, but &lt;/span&gt;&lt;strong&gt;&lt;span&gt;real buyer journeys are a sequence&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;. Most people prompt and then have follow-up questions based on the AI’s answer.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;So my script runs an opening question and its follow-ups in the same chat, saving the data at each step. While a single prompt is still a good way to check whether a brand shows up, the follow-ups show what ChatGPT goes looking for once someone starts deciding.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Now my script was ready for testing.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;My Test: Three Conversations, Three Turns Each&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;On October 1, I ran three conversations that I wrote for this test (I chose random topics). Each one happened in a single chat, so every follow-up built on the answer before it. The follow-ups are my guesses at what someone would ask next. &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;All nine answers came from one run on the free plan, and the stream showed ChatGPT's smaller "mini" model on every turn.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;ChatGPT’s Initial Queries Were More Specific Than My Prompts and Even Listed Brands&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT rewrites your question into its own search queries, and the stream shows them. Here's the first query in each conversation, alongside my original ask and what I was planning to ask for turn number two.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In the first turn, &lt;/span&gt;&lt;strong&gt;&lt;span&gt;ChatGPT got more specific in its queries than any of the prompts I used, even including specific brands&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;. It also started querying topics and features that I was planning to ask in the second turn.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The luggage query surprised me most. I never mentioned wheels or handles in the first question, yet "durable wheels handles" was in the very first search, one turn before I planned to ask about those features. The cloud backup query did something similar, adding "pricing" a turn before I asked about cost. &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I wrote those follow-ups because they seemed like the obvious next questions, and ChatGPT apparently made a similar guess. I can’t tell where that comes from (training data, my wording, or something else) based on the query fan-outs. But the query does give me a written list of what ChatGPT thinks this buyer cares about.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;How Did ChatGPT Get to Its Answer?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;There was one question I kept coming back to: how did ChatGPT decide which product to recommend? &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I used the suitcase conversation as the example here. I wanted to understand why ChatGPT picked the Travelpro Maxlite 5 (its final recommendation), and how it decided that was the most durable option.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;&lt;strong&gt;Turn 1: ChatGPT Recommends Its Preferred Product&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;ChatGPT first brought up the Travelpro Maxlite 5 in its response to turn one:&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;And the initial reasons it gave (turn one) cited Pack &amp;amp; Porter and mentioned price, weight, and value:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The stream doesn't show ChatGPT's reasoning, but it does show what was on the table. ChatGPT ranked the Maxlite 5 first and backed it with two Pack &amp;amp; Porter pages. I read the Pack &amp;amp; Porter's "Best Carry-On Luggage Under $200" page, which ranked the Maxlite 5 first and carried the specs from ChatGPT’s answer:&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;5.4 lb&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;A lifetime warranty&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Published dimensions of 23 × 14.5 × 9 inches, which is where ChatGPT's warning about strict 22 × 14 × 9 airline limits came from&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;For the Samsonite Freeform: 6.5 lb, a polypropylene shell, and a 10-year warranty&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;So the recommendation lines up with the top-cited site. That doesn’t mean Pack &amp;amp; Porter caused the pick: the brand list was already in the query before any page came back.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;Turns 2 and 3: ChatGPT Looks to Validate Its Choice&lt;strong&gt;&lt;span style="white-space-collapse: preserve;"&gt; &lt;/span&gt;&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Turn two is where things got interesting. When I asked which suitcase it would choose for durable wheels and handles, ChatGPT ran exactly one search: &lt;span style="color: #188038;"&gt;Travelpro Maxlite 5 wheels handle warranty durable review&lt;/span&gt;&lt;/span&gt;&lt;span style="white-space-collapse: preserve;"&gt;. &lt;/span&gt;&lt;strong&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;br&gt;&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Instead of looking up which carry-on is most durable, ChatGPT looked for support for the product it had already picked. &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Whatever drove the choice happened before that search. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The sources cited in turn two both have reviews that describe hands-on testing. GearLab ran drop tests (2022). Pack Hacker logged a two-week trip and wrote that the handle "does wiggle substantially" when extended, which shows up in the answer as "can feel somewhat wobbly." The supporting "+1" source, NewCarryOn's Platinum Elite review, relied on OutdoorGearLab's tests rather than its own.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;There were two 2026 reviews from the same source set that didn’t get cited, and both included on-page disclaimers that they never tested the bag. A third 2026 review that claimed a year of testing wasn't cited either, so I'm treating this as one data point.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;By the third turn on warranty and repair, the search covered a single site.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; All 13 results came from travelpro.com, and the summary line read "Searched 1 website."&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;It’s also worth noting that in the data, the first turn's sources already included a Pack &amp;amp; Porter page that ranked carry-ons by warranty.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;A Carry-On Question Returned a Checked-Bag Record&lt;/h3&gt; 
&lt;p&gt;One odd thing I noted from the carry-on convo: a product recommended by ChatGPT showed pricing/product disparities. The American Tourister product card showed $174.99, while the answer’s table said about $72. The stream included offers at both prices, but the $174.99 brand offer linked to a large checked bag - not a carry-on. For ecommerce brands, I’d check that the product records match the variant being recommended.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;How Early in the Conversation Did ChatGPT Start Citing a Brand Directly?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Turn one of the luggage conversation cited a review site. By the third turn, every citation was Travelpro's own site:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Only the luggage conversation shifted from third-party reviews to the brand. The cloud backup and Spanish conversations went to brand sites from the first turn (and featured a lot of help-center and policy pages): &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;All 36 citations across those two chats were on the brands' own sites&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Half of all the queries in the run (12 of 24) included the word "official"&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Two queries used &lt;/span&gt;&lt;span style="color: #188038;"&gt;site:&lt;/span&gt;&lt;span&gt; to search a single domain&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;The opening Spanish question did retrieve ten third-party "best Spanish apps" roundups, but all ten appeared only in the second set of results.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If I think about it as a funnel, it makes sense that ChatGPT often turned directly to the brands to answer product-spec questions (such as warranty, cost, and cancellation policies). &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Even if your brand isn't cited early in a conversation, you should have help articles, warranty pages, and pricing pages ready for the questions people ask right before they buy.&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;Every Cited URL Appeared in the First Result Set&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;The stream recorded an initial set of search results and additional results later. Across all eight turns that searched, every cited URL appeared in the first set. None of the results that appeared only later were cited. I don’t know why, but it gives us another thing to investigate: did our page appear in the set containing the cited sources, or only among the additional results?&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here's the first set for the opening luggage question, in the order the stream numbered it:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Interestingly, &lt;strong&gt;five of the nine citations came from one travel gear review site&lt;/strong&gt;: packandporter.com. This implies that ChatGPT places a high degree of trust in content from Pack &amp;amp; Porter, perhaps because the site explicitly states who writes the reviews, how they rank products, and even how they earn money. &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The page titles caught my eye too. The two cited pages are the closest match to what ChatGPT searched for: &lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Carry-on&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Under $200&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Travelpro vs Samsonite&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;A checked-luggage list and a general "luggage under $200" list came back in the same batch and weren't cited. Considering this is a small data set, I'm treating title match as something to test further. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you track AI visibility, this is worth checking for your own pages. A page can be retrieved in either set, and in this run only one set produced citations.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;Where ChatGPT's Answers Diverged from Sources&lt;/span&gt;&lt;/h2&gt; 
&lt;h3&gt;&lt;span&gt;It Rated Duolingo Without a Corresponding Citation&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Duolingo was in the first Spanish query, but ChatGPT never ran a Duolingo-specific search, and no duolingo.com page came back.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The answer still gave Duolingo two stars for speaking practice, and it’s unclear where that rating came from. It could be training data, the conversation so far, or a third-party page that mentions Duolingo. Duolingo does document a speaking feature on its blog, Video Call, which is part of its Max plan, but that page didn't appear in the results.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Surprisingly, ChatGPT's suggested follow-up dropped Duolingo. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;br&gt;On the next turn (turn three), no new search was recorded at all. It picked Pimsleur and started answering in under two seconds.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;strong&gt;&lt;span&gt;ChatGPT Ignored Prices That Loaded with JavaScript &lt;/span&gt;&lt;/strong&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;The most interesting finding that I want to run further tests on involved how pricing information was listed on a brand’s site. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In the Spanish chat's pricing turn (turn three), the brand with a price in their page's HTML was quoted exactly, while two others had "pricing varies" language.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I checked the source of each pricing page to see what each brand’s pages said and how they were formatted. I did this on October 2, so this information is not necessarily what ChatGPT received:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Pimsleur:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The page's HTML includes structured subscription data, "Monthly 7-day free trial" at 19.95, even though the visible text rounds to "$20." The answer quoted "$19.95/month" and the 7-day trial.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Babbel:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The prices page loads with almost nothing in it until JavaScript runs ("JavaScript required to see this page"). The answer still quoted Babbel's annual plan ($8.95/month equivalent, $107.40 billed upfront), but said its monthly pricing "varies by offer." Monthly pricing was what I'd asked about.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Busuu:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The plans page had no prices in its HTML. The answer said pricing "varies by plan/platform."&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;There's outside support for why this could happen. &lt;/span&gt;&lt;a href="https://vercel.com/blog/the-rise-of-the-ai-crawler"&gt;&lt;u&gt;&lt;span&gt;Vercel's 2024 study of AI crawlers&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;strong&gt;&lt;span&gt;found that OpenAI's crawlers download JavaScript files but don't run them&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;, so content that only appears after JavaScript runs may never reach them.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Babbel's annual price still got through from somewhere, so JavaScript content isn't simply invisible. And ChatGPT had already picked Pimsleur a turn earlier without searching, so I can't credit the pricing page for the pick. We've been testing server-rendered HTML against JavaScript-rendered content for clients, and this gives us a clean side-by-side to build on.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;strong&gt;&lt;span&gt;CrashPlan's Own Pages Disagreed&lt;/span&gt;&lt;/strong&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Cloud backup had a pricing problem of a different kind. When I asked about cost, ChatGPT searched &lt;/span&gt;&lt;span style="color: #188038;"&gt;site:crashplan.com personal backup discontinued individual 2026&lt;/span&gt;&lt;span&gt;, checking whether CrashPlan still sells a personal plan. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;CrashPlan's help center says its Essential plan is a legacy plan that's no longer available for new purchase. Meanwhile, a CrashPlan Essential page is still live, offering $2.99 a month and a "Start Free Trial" button. ChatGPT quoted my theoretical user, a solo photographer, with the business rate of $7.33 per user per month, and told me to ignore "the older $2.99/month Essential price you may see elsewhere."&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;For me, this raises a straightforward question: &lt;/span&gt;&lt;strong&gt;&lt;span&gt;are your pricing pages and help center describing the same offer?&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; If a plan is no longer available to new customers, its old landing page shouldn’t still make it look available. I’d check those pages together, clear up any conflicting information, and rerun the question to see whether the answer changes.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;What I'd Test Based on These Findings&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;None of these have been run yet. Each one comes straight from something in the data above.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;You don’t need a script to start digging. Open ChatGPT in Chrome, open DevTools on the Network tab, and ask your category’s main question. Find the response for that conversation and search it for the query ChatGPT wrote. Then ask your two most likely follow-ups in the same chat and note how the queries and cited sites change.&lt;/span&gt;&lt;/p&gt; 
&lt;ol&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;What did ChatGPT add to the first query?&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Note what it added on its own, such as “durable wheels handles.” Check whether your page answers that concern in plain text. If it doesn’t, add it. Then rerun the same conversation over several weeks and track which set of results your page lands in, as well as whether it gets cited.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Do your help pages answer the follow-up questions? &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Write down the likely next question in your category, like a warranty claim, a cancellation, or recovering a deleted file. Check that a help article answers it directly, then compare what ChatGPT says against the actual policy.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Does putting the price in HTML change the answer? &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Take a page where pricing only appears after JavaScript runs. Add plain-text pricing and cancellation terms to the HTML, and leave a similar page unchanged. Then measure whether ChatGPT quotes the exact price or says pricing “varies.”&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Which brands does ChatGPT search for?&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Ask the opening question in fresh chats several times. Record which brands appear in the first query, which get a follow-up query of their own, and which get described without any of their own pages in the results.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Do the product records match the recommendation?&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; For shopping categories, compare the product records ChatGPT carries against the product the answer is recommending. A carry-on question carrying a checked-bag record is the kind of mismatch to look for.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Do cited page titles match the query? &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;For the first set in your category, compare each page’s title to ChatGPT’s query. Over repeated runs, check whether the cited pages are consistently the closest title match. If they are, test retitling a page to match the query wording and see whether it starts getting cited.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ol&gt; 
&lt;p&gt;&lt;span&gt;Answers vary from run to run, so these tests need repeated captures before and after a change before I’d call something an improvement. Save what you capture so you can check it again later.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;Things I Still Don't Know&lt;/span&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Whether this repeats.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; This was one run per conversation, on the mini model, in my own account. The sidebar showed earlier chats on these topics from a first attempt, so I can't rule out account history.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;What the model had while writing.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The stream shows when results reached the browser, not what was available to ChatGPT behind the scenes.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Which search engine supplies the results.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; OpenAI says ChatGPT search "sometimes partners with other search providers." &lt;/span&gt;&lt;a href="https://docs.peec.ai/research/chatgpt-search-result-providers.md"&gt;&lt;u&gt;&lt;span&gt;Peec's research&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; maps several providers, including Google results gathered by third parties and Microsoft's grounding service. I ran four of the captured queries in Google and Bing myself a day later, from my own browser, so treat this as a rough check: only 3 of 32 Google results and 2 of 19 Bing results matched ChatGPT's first set. Ranking on Google wasn't a good stand-in for being in that set here.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Why a page lands in the first set or the second,&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; and whether being in the first set means the page was read in full.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Whether Pack &amp;amp; Porter's ranking drove the recommendation,&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; or whether both reflect what ChatGPT already believed about these bags.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Where the brand shortlist comes from.&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;I built this tool to better support the questions our clients ask. Knowing whether a brand shows up in the first answer is useful. Seeing what ChatGPT goes looking for when someone pushes back on that answer gives us much more to work with.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I&lt;/span&gt;&lt;span&gt;f you want to talk through what this could mean for your AI search strategy, &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/contact"&gt;&lt;u&gt;&lt;span&gt;chat with the Seer team&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt;.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;em&gt;&lt;span&gt;Closing note:&lt;/span&gt;&lt;/em&gt;&lt;/strong&gt;&lt;em&gt;&lt;span&gt; This is a tracker I'm building as I go, and what ChatGPT exposes keeps shifting. I keep records of the raw streams so I can go back and recheck patterns like the citation split against older captures. I can also monitor the parts we don't understand yet, and rerun fixed prompts to see what changes.&lt;/span&gt;&lt;/em&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;strong&gt;&lt;span&gt;Methodology&lt;/span&gt;&lt;/strong&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;I captured three conversations on October 1, 2026, each with three prewritten prompts asked in the same chat, for nine prompts in total. They ran on the ChatGPT free plan in a desktop Chrome browser, and every turn was answered by the mini model. For each turn, we saved the answer, a screenshot and the full stream. Counts come from the stream itself. Pages were fetched on October 2, the day after the capture, and may differ from what was live during the run. None of the proposed tests have been run.&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-search-tool" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/ChatGPT-search-tool-Hoa-blog.png" alt="I Built a Tool to Watch ChatGPT Search, and It Predicted My Next Question" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;I wanted to understand how ChatGPT searched, so I wrote a script to watch it and tracked 3 random shopping conversations: a carry-on suitcase, a cloud backup tool, and a Spanish app&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;For each topic, I asked an opening question plus two follow-ups in the same chat and saved the data as ChatGPT was building the answer&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;In 2 of the 3 conversations, ChatGPT's first search added a concern I hadn't raised yet but planned to ask about&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;ChatGPT started naming brands during its first search and cited brand sites in that first response for 2 of the 3 conversations&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;In the luggage conversation, ChatGPT cited one review site for 5 of the 9 sources on the first question, and the recommendation lined up with that site's ranking - then ChatGPT looked for further evidence to support that pick in subsequent answers&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;We can use this tool to help clients understand if their brand shows up in the first answer, and what ChatGPT goes looking for when someone pushes back on that answer&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt;  
&lt;h2&gt;&lt;span&gt;De-Mystifying the Black Box of AI Search&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;AI search is a black box. We still don’t know a lot about how it searches, where it pulls information from, or why it cites some sources but not others. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;After &lt;/span&gt;&lt;a href="https://suganthan.com/blog/chatgpt-new-search-language/"&gt;&lt;u&gt;&lt;span&gt;Suganthan Mohanadasan's blog on ChatGPT search language&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; made the rounds in our internal Slack, I was inspired to figure out how to reverse-engineer ChatGPT’s AI search. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I wanted a diagnostic engine that would simulate realistic user conversations and capture the complete retrieval layer, so we could monitor the AI prompts and conversations our clients care about. So I built one.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Instead of just knowing whether a brand shows up in a response, I can see exactly what ChatGPT searches for.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Our team can use this information to see which sites ChatGPT queries to find information about a client’s vertical, analyze why a competitor was cited instead of a client, and even run tactical tests on client content.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Let’s dive into what I built and my initial findings. &lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;I Built A Script to Watch ChatGPT Search in Real Time&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Suganathan’s blog looked at a single query, but I wanted the backend data. Basically, the AI has its own language it speaks when fetching real-time data, and you can see it if you know where to look.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;While ChatGPT builds an answer, the website streams data to your browser, including: &lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;What searches it runs&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;The results that come back &lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Which of those become citations &lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Product records &lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;I'll call that information the stream. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Working with AI coding tools, &lt;/span&gt;&lt;strong&gt;&lt;span&gt;I built a script that types prompts into ChatGPT's website the way a person would&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;. It saves the answer and a screenshot, and it keeps the full stream for every turn.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The other challenge I had to address was Suganthan’s post looked at a single query, but &lt;/span&gt;&lt;strong&gt;&lt;span&gt;real buyer journeys are a sequence&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;. Most people prompt and then have follow-up questions based on the AI’s answer.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;So my script runs an opening question and its follow-ups in the same chat, saving the data at each step. While a single prompt is still a good way to check whether a brand shows up, the follow-ups show what ChatGPT goes looking for once someone starts deciding.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Now my script was ready for testing.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;My Test: Three Conversations, Three Turns Each&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;On October 1, I ran three conversations that I wrote for this test (I chose random topics). Each one happened in a single chat, so every follow-up built on the answer before it. The follow-ups are my guesses at what someone would ask next. &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;All nine answers came from one run on the free plan, and the stream showed ChatGPT's smaller "mini" model on every turn.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;ChatGPT’s Initial Queries Were More Specific Than My Prompts and Even Listed Brands&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;ChatGPT rewrites your question into its own search queries, and the stream shows them. Here's the first query in each conversation, alongside my original ask and what I was planning to ask for turn number two.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In the first turn, &lt;/span&gt;&lt;strong&gt;&lt;span&gt;ChatGPT got more specific in its queries than any of the prompts I used, even including specific brands&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;. It also started querying topics and features that I was planning to ask in the second turn.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The luggage query surprised me most. I never mentioned wheels or handles in the first question, yet "durable wheels handles" was in the very first search, one turn before I planned to ask about those features. The cloud backup query did something similar, adding "pricing" a turn before I asked about cost. &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I wrote those follow-ups because they seemed like the obvious next questions, and ChatGPT apparently made a similar guess. I can’t tell where that comes from (training data, my wording, or something else) based on the query fan-outs. But the query does give me a written list of what ChatGPT thinks this buyer cares about.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;How Did ChatGPT Get to Its Answer?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;There was one question I kept coming back to: how did ChatGPT decide which product to recommend? &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I used the suitcase conversation as the example here. I wanted to understand why ChatGPT picked the Travelpro Maxlite 5 (its final recommendation), and how it decided that was the most durable option.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;&lt;strong&gt;Turn 1: ChatGPT Recommends Its Preferred Product&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;ChatGPT first brought up the Travelpro Maxlite 5 in its response to turn one:&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;And the initial reasons it gave (turn one) cited Pack &amp;amp; Porter and mentioned price, weight, and value:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The stream doesn't show ChatGPT's reasoning, but it does show what was on the table. ChatGPT ranked the Maxlite 5 first and backed it with two Pack &amp;amp; Porter pages. I read the Pack &amp;amp; Porter's "Best Carry-On Luggage Under $200" page, which ranked the Maxlite 5 first and carried the specs from ChatGPT’s answer:&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;5.4 lb&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;A lifetime warranty&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Published dimensions of 23 × 14.5 × 9 inches, which is where ChatGPT's warning about strict 22 × 14 × 9 airline limits came from&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;For the Samsonite Freeform: 6.5 lb, a polypropylene shell, and a 10-year warranty&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;So the recommendation lines up with the top-cited site. That doesn’t mean Pack &amp;amp; Porter caused the pick: the brand list was already in the query before any page came back.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;span&gt;Turns 2 and 3: ChatGPT Looks to Validate Its Choice&lt;strong&gt;&lt;span style="white-space-collapse: preserve;"&gt; &lt;/span&gt;&lt;/strong&gt;&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Turn two is where things got interesting. When I asked which suitcase it would choose for durable wheels and handles, ChatGPT ran exactly one search: &lt;span style="color: #188038;"&gt;Travelpro Maxlite 5 wheels handle warranty durable review&lt;/span&gt;&lt;/span&gt;&lt;span style="white-space-collapse: preserve;"&gt;. &lt;/span&gt;&lt;strong&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;br&gt;&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Instead of looking up which carry-on is most durable, ChatGPT looked for support for the product it had already picked. &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Whatever drove the choice happened before that search. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The sources cited in turn two both have reviews that describe hands-on testing. GearLab ran drop tests (2022). Pack Hacker logged a two-week trip and wrote that the handle "does wiggle substantially" when extended, which shows up in the answer as "can feel somewhat wobbly." The supporting "+1" source, NewCarryOn's Platinum Elite review, relied on OutdoorGearLab's tests rather than its own.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;There were two 2026 reviews from the same source set that didn’t get cited, and both included on-page disclaimers that they never tested the bag. A third 2026 review that claimed a year of testing wasn't cited either, so I'm treating this as one data point.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;By the third turn on warranty and repair, the search covered a single site.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; All 13 results came from travelpro.com, and the summary line read "Searched 1 website."&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;It’s also worth noting that in the data, the first turn's sources already included a Pack &amp;amp; Porter page that ranked carry-ons by warranty.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;A Carry-On Question Returned a Checked-Bag Record&lt;/h3&gt; 
&lt;p&gt;One odd thing I noted from the carry-on convo: a product recommended by ChatGPT showed pricing/product disparities. The American Tourister product card showed $174.99, while the answer’s table said about $72. The stream included offers at both prices, but the $174.99 brand offer linked to a large checked bag - not a carry-on. For ecommerce brands, I’d check that the product records match the variant being recommended.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;How Early in the Conversation Did ChatGPT Start Citing a Brand Directly?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Turn one of the luggage conversation cited a review site. By the third turn, every citation was Travelpro's own site:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Only the luggage conversation shifted from third-party reviews to the brand. The cloud backup and Spanish conversations went to brand sites from the first turn (and featured a lot of help-center and policy pages): &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;All 36 citations across those two chats were on the brands' own sites&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Half of all the queries in the run (12 of 24) included the word "official"&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Two queries used &lt;/span&gt;&lt;span style="color: #188038;"&gt;site:&lt;/span&gt;&lt;span&gt; to search a single domain&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;The opening Spanish question did retrieve ten third-party "best Spanish apps" roundups, but all ten appeared only in the second set of results.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If I think about it as a funnel, it makes sense that ChatGPT often turned directly to the brands to answer product-spec questions (such as warranty, cost, and cancellation policies). &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Even if your brand isn't cited early in a conversation, you should have help articles, warranty pages, and pricing pages ready for the questions people ask right before they buy.&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;Every Cited URL Appeared in the First Result Set&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;The stream recorded an initial set of search results and additional results later. Across all eight turns that searched, every cited URL appeared in the first set. None of the results that appeared only later were cited. I don’t know why, but it gives us another thing to investigate: did our page appear in the set containing the cited sources, or only among the additional results?&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here's the first set for the opening luggage question, in the order the stream numbered it:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Interestingly, &lt;strong&gt;five of the nine citations came from one travel gear review site&lt;/strong&gt;: packandporter.com. This implies that ChatGPT places a high degree of trust in content from Pack &amp;amp; Porter, perhaps because the site explicitly states who writes the reviews, how they rank products, and even how they earn money. &lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The page titles caught my eye too. The two cited pages are the closest match to what ChatGPT searched for: &lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Carry-on&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Under $200&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Travelpro vs Samsonite&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;A checked-luggage list and a general "luggage under $200" list came back in the same batch and weren't cited. Considering this is a small data set, I'm treating title match as something to test further. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you track AI visibility, this is worth checking for your own pages. A page can be retrieved in either set, and in this run only one set produced citations.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;Where ChatGPT's Answers Diverged from Sources&lt;/span&gt;&lt;/h2&gt; 
&lt;h3&gt;&lt;span&gt;It Rated Duolingo Without a Corresponding Citation&lt;/span&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Duolingo was in the first Spanish query, but ChatGPT never ran a Duolingo-specific search, and no duolingo.com page came back.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The answer still gave Duolingo two stars for speaking practice, and it’s unclear where that rating came from. It could be training data, the conversation so far, or a third-party page that mentions Duolingo. Duolingo does document a speaking feature on its blog, Video Call, which is part of its Max plan, but that page didn't appear in the results.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Surprisingly, ChatGPT's suggested follow-up dropped Duolingo. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;br&gt;On the next turn (turn three), no new search was recorded at all. It picked Pimsleur and started answering in under two seconds.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;strong&gt;&lt;span&gt;ChatGPT Ignored Prices That Loaded with JavaScript &lt;/span&gt;&lt;/strong&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;The most interesting finding that I want to run further tests on involved how pricing information was listed on a brand’s site. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;In the Spanish chat's pricing turn (turn three), the brand with a price in their page's HTML was quoted exactly, while two others had "pricing varies" language.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I checked the source of each pricing page to see what each brand’s pages said and how they were formatted. I did this on October 2, so this information is not necessarily what ChatGPT received:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Pimsleur:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The page's HTML includes structured subscription data, "Monthly 7-day free trial" at 19.95, even though the visible text rounds to "$20." The answer quoted "$19.95/month" and the 7-day trial.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Babbel:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The prices page loads with almost nothing in it until JavaScript runs ("JavaScript required to see this page"). The answer still quoted Babbel's annual plan ($8.95/month equivalent, $107.40 billed upfront), but said its monthly pricing "varies by offer." Monthly pricing was what I'd asked about.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Busuu:&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The plans page had no prices in its HTML. The answer said pricing "varies by plan/platform."&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;br&gt; 
&lt;p&gt;&lt;span&gt;There's outside support for why this could happen. &lt;/span&gt;&lt;a href="https://vercel.com/blog/the-rise-of-the-ai-crawler"&gt;&lt;u&gt;&lt;span&gt;Vercel's 2024 study of AI crawlers&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;strong&gt;&lt;span&gt;found that OpenAI's crawlers download JavaScript files but don't run them&lt;/span&gt;&lt;/strong&gt;&lt;span&gt;, so content that only appears after JavaScript runs may never reach them.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Babbel's annual price still got through from somewhere, so JavaScript content isn't simply invisible. And ChatGPT had already picked Pimsleur a turn earlier without searching, so I can't credit the pricing page for the pick. We've been testing server-rendered HTML against JavaScript-rendered content for clients, and this gives us a clean side-by-side to build on.&lt;/span&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;strong&gt;&lt;span&gt;CrashPlan's Own Pages Disagreed&lt;/span&gt;&lt;/strong&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;Cloud backup had a pricing problem of a different kind. When I asked about cost, ChatGPT searched &lt;/span&gt;&lt;span style="color: #188038;"&gt;site:crashplan.com personal backup discontinued individual 2026&lt;/span&gt;&lt;span&gt;, checking whether CrashPlan still sells a personal plan. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;CrashPlan's help center says its Essential plan is a legacy plan that's no longer available for new purchase. Meanwhile, a CrashPlan Essential page is still live, offering $2.99 a month and a "Start Free Trial" button. ChatGPT quoted my theoretical user, a solo photographer, with the business rate of $7.33 per user per month, and told me to ignore "the older $2.99/month Essential price you may see elsewhere."&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;For me, this raises a straightforward question: &lt;/span&gt;&lt;strong&gt;&lt;span&gt;are your pricing pages and help center describing the same offer?&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; If a plan is no longer available to new customers, its old landing page shouldn’t still make it look available. I’d check those pages together, clear up any conflicting information, and rerun the question to see whether the answer changes.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;What I'd Test Based on These Findings&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;None of these have been run yet. Each one comes straight from something in the data above.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;You don’t need a script to start digging. Open ChatGPT in Chrome, open DevTools on the Network tab, and ask your category’s main question. Find the response for that conversation and search it for the query ChatGPT wrote. Then ask your two most likely follow-ups in the same chat and note how the queries and cited sites change.&lt;/span&gt;&lt;/p&gt; 
&lt;ol&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;What did ChatGPT add to the first query?&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Note what it added on its own, such as “durable wheels handles.” Check whether your page answers that concern in plain text. If it doesn’t, add it. Then rerun the same conversation over several weeks and track which set of results your page lands in, as well as whether it gets cited.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Do your help pages answer the follow-up questions? &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Write down the likely next question in your category, like a warranty claim, a cancellation, or recovering a deleted file. Check that a help article answers it directly, then compare what ChatGPT says against the actual policy.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Does putting the price in HTML change the answer? &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Take a page where pricing only appears after JavaScript runs. Add plain-text pricing and cancellation terms to the HTML, and leave a similar page unchanged. Then measure whether ChatGPT quotes the exact price or says pricing “varies.”&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Which brands does ChatGPT search for?&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Ask the opening question in fresh chats several times. Record which brands appear in the first query, which get a follow-up query of their own, and which get described without any of their own pages in the results.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Do the product records match the recommendation?&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; For shopping categories, compare the product records ChatGPT carries against the product the answer is recommending. A carry-on question carrying a checked-bag record is the kind of mismatch to look for.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Do cited page titles match the query? &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;For the first set in your category, compare each page’s title to ChatGPT’s query. Over repeated runs, check whether the cited pages are consistently the closest title match. If they are, test retitling a page to match the query wording and see whether it starts getting cited.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ol&gt; 
&lt;p&gt;&lt;span&gt;Answers vary from run to run, so these tests need repeated captures before and after a change before I’d call something an improvement. Save what you capture so you can check it again later.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;Things I Still Don't Know&lt;/span&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Whether this repeats.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; This was one run per conversation, on the mini model, in my own account. The sidebar showed earlier chats on these topics from a first attempt, so I can't rule out account history.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;What the model had while writing.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; The stream shows when results reached the browser, not what was available to ChatGPT behind the scenes.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Which search engine supplies the results.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; OpenAI says ChatGPT search "sometimes partners with other search providers." &lt;/span&gt;&lt;a href="https://docs.peec.ai/research/chatgpt-search-result-providers.md"&gt;&lt;u&gt;&lt;span&gt;Peec's research&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; maps several providers, including Google results gathered by third parties and Microsoft's grounding service. I ran four of the captured queries in Google and Bing myself a day later, from my own browser, so treat this as a rough check: only 3 of 32 Google results and 2 of 19 Bing results matched ChatGPT's first set. Ranking on Google wasn't a good stand-in for being in that set here.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Why a page lands in the first set or the second,&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; and whether being in the first set means the page was read in full.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Whether Pack &amp;amp; Porter's ranking drove the recommendation,&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; or whether both reflect what ChatGPT already believed about these bags.&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;strong&gt;&lt;span&gt;Where the brand shortlist comes from.&lt;/span&gt;&lt;/strong&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;I built this tool to better support the questions our clients ask. Knowing whether a brand shows up in the first answer is useful. Seeing what ChatGPT goes looking for when someone pushes back on that answer gives us much more to work with.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I&lt;/span&gt;&lt;span&gt;f you want to talk through what this could mean for your AI search strategy, &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/contact"&gt;&lt;u&gt;&lt;span&gt;chat with the Seer team&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt;.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;em&gt;&lt;span&gt;Closing note:&lt;/span&gt;&lt;/em&gt;&lt;/strong&gt;&lt;em&gt;&lt;span&gt; This is a tracker I'm building as I go, and what ChatGPT exposes keeps shifting. I keep records of the raw streams so I can go back and recheck patterns like the citation split against older captures. I can also monitor the parts we don't understand yet, and rerun fixed prompts to see what changes.&lt;/span&gt;&lt;/em&gt;&lt;/p&gt; 
&lt;h3&gt;&lt;strong&gt;&lt;span&gt;Methodology&lt;/span&gt;&lt;/strong&gt;&lt;/h3&gt; 
&lt;p&gt;&lt;span&gt;I captured three conversations on October 1, 2026, each with three prewritten prompts asked in the same chat, for nine prompts in total. They ran on the ChatGPT free plan in a desktop Chrome browser, and every turn was answered by the mini model. For each turn, we saved the answer, a screenshot and the full stream. Counts come from the stream itself. Pages were fetched on October 2, the day after the capture, and may differ from what was live during the run. None of the proposed tests have been run.&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fchatgpt-search-tool&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Author: Hoa Cong</category>
      <category>AI</category>
      <pubDate>Fri, 30 Oct 2026 17:04:41 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/chatgpt-search-tool</guid>
      <dc:date>2026-10-30T17:04:41Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>Agentic Commerce May Break Your Measurement Before It Breaks Your Strategy.</title>
      <link>https://www.seerinteractive.com/insights/agent-owned-checkouts</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/agent-owned-checkouts" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Blog%20Update%20%231%20(9).png" alt="Agentic Commerce May Break Your Measurement Before It Breaks Your Strategy." class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;p&gt;AI agents can complete a purchase without ever loading your site, so GA4 may never record the sale while your order system does.&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/agent-owned-checkouts" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Blog%20Update%20%231%20(9).png" alt="Agentic Commerce May Break Your Measurement Before It Breaks Your Strategy." class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;p&gt;AI agents can complete a purchase without ever loading your site, so GA4 may never record the sale while your order system does.&lt;/p&gt;   
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fagent-owned-checkouts&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Analytics</category>
      <category>AI</category>
      <category>Agentic</category>
      <category>Travel &amp; Hospitality</category>
      <category>Author: Corey Abramson</category>
      <pubDate>Wed, 28 Oct 2026 19:01:08 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/agent-owned-checkouts</guid>
      <dc:date>2026-10-28T19:01:08Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>Same Prompt, Different Data Structure: ChatGPT's Search Data Keeps Changing</title>
      <link>https://www.seerinteractive.com/insights/chatgpt-data-structure</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-data-structure" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Data%20Structure%20blog%20(1).png" alt="Same Prompt, Different Data Structure: ChatGPT's Search Data Keeps Changing" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;ul&gt; 
 &lt;li&gt;I ran the same ChatGPT prompt multiple times and saw the search data change across weeks, days, and even minutes&lt;/li&gt; 
 &lt;li&gt;Between Sept. 3 and Sept. 16, ChatGPT's underlying pipe-language data dropped from 4 lines to 0, while unique retrieved URLs went up (from 14 to 40)&lt;/li&gt; 
 &lt;li&gt;Two identical runs 8 minutes apart on Sept. 17 shared zero exact query strings - the same prompt produced different search behavior&lt;/li&gt; 
 &lt;li&gt;This tells us that measuring AI search visibility through single-run prompts is misleading and ineffective&lt;/li&gt; 
 &lt;li&gt;It also shows the value of looking at the full retrieval process (not just the rendered answer) when reporting on AI brand visibility: 34 of the 40 URLs retrieved on Sept 16 were retrieved but not cited&lt;/li&gt; 
&lt;/ul&gt;  
&lt;p&gt;&lt;a href="#methodology"&gt;Skip to the methodology →&lt;/a&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/chatgpt-data-structure" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Data%20Structure%20blog%20(1).png" alt="Same Prompt, Different Data Structure: ChatGPT's Search Data Keeps Changing" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt;  TL;DR 
&lt;ul&gt; 
 &lt;li&gt;I ran the same ChatGPT prompt multiple times and saw the search data change across weeks, days, and even minutes&lt;/li&gt; 
 &lt;li&gt;Between Sept. 3 and Sept. 16, ChatGPT's underlying pipe-language data dropped from 4 lines to 0, while unique retrieved URLs went up (from 14 to 40)&lt;/li&gt; 
 &lt;li&gt;Two identical runs 8 minutes apart on Sept. 17 shared zero exact query strings - the same prompt produced different search behavior&lt;/li&gt; 
 &lt;li&gt;This tells us that measuring AI search visibility through single-run prompts is misleading and ineffective&lt;/li&gt; 
 &lt;li&gt;It also shows the value of looking at the full retrieval process (not just the rendered answer) when reporting on AI brand visibility: 34 of the 40 URLs retrieved on Sept 16 were retrieved but not cited&lt;/li&gt; 
&lt;/ul&gt;  
&lt;p&gt;&lt;a href="#methodology"&gt;Skip to the methodology →&lt;/a&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fchatgpt-data-structure&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Author: Hoa Cong</category>
      <category>AI</category>
      <pubDate>Thu, 22 Oct 2026 19:00:00 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/chatgpt-data-structure</guid>
      <dc:date>2026-10-22T19:00:00Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>Competitive Analysis for Paid Media: How To Find the Broken Stage</title>
      <link>https://www.seerinteractive.com/insights/competitive-analysis-for-paid-media-find-the-broken-stage</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/competitive-analysis-for-paid-media-find-the-broken-stage" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Market%20vs%20operational%20blog%20(1).png" alt="Competitive Analysis for Paid Media: How To Find the Broken Stage" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;Competitive research can show which stage of your funnel is broken (Be Seen, Be Believed, or Be Chosen) and which channel to fix it in.&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/competitive-analysis-for-paid-media-find-the-broken-stage" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Market%20vs%20operational%20blog%20(1).png" alt="Competitive Analysis for Paid Media: How To Find the Broken Stage" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;Competitive research can show which stage of your funnel is broken (Be Seen, Be Believed, or Be Chosen) and which channel to fix it in.&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fcompetitive-analysis-for-paid-media-find-the-broken-stage&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Paid Media</category>
      <category>AI</category>
      <category>Author: Brittany Sager</category>
      <category>chatgpt</category>
      <category>ad strategy</category>
      <pubDate>Wed, 07 Oct 2026 19:01:13 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/competitive-analysis-for-paid-media-find-the-broken-stage</guid>
      <dc:date>2026-10-07T19:01:13Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>Re-imagining AI Visibility KPIs in a Query Fan Out world</title>
      <link>https://www.seerinteractive.com/insights/ai-visibility-kpis</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/ai-visibility-kpis" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Wil%20Blog%20OG%20(7).png" alt="Re-imagining AI Visibility KPIs in a Query Fan Out world" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;p&gt;&lt;a href="#KPIs"&gt;&lt;span&gt;Wil, take me right to the KPI for long term sustainable GEO visibility!!&lt;/span&gt;&lt;/a&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/ai-visibility-kpis" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Wil%20Blog%20OG%20(7).png" alt="Re-imagining AI Visibility KPIs in a Query Fan Out world" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;p&gt;&lt;a href="#KPIs"&gt;&lt;span&gt;Wil, take me right to the KPI for long term sustainable GEO visibility!!&lt;/span&gt;&lt;/a&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fai-visibility-kpis&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>author-wil-reynolds</category>
      <category>AI</category>
      <pubDate>Fri, 02 Oct 2026 16:30:48 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/ai-visibility-kpis</guid>
      <dc:date>2026-10-02T16:30:48Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>SaaS Marketers - Promote your Claude &amp; Codex connectors &amp; win in AI search</title>
      <link>https://www.seerinteractive.com/insights/saas-marketers-promote-your-claude-or-codex-connectors</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/saas-marketers-promote-your-claude-or-codex-connectors" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Wil%20Blog%20OG%20(6)-png.png" alt="SaaS Marketers - Promote your Claude &amp;amp; Codex connectors &amp;amp; win in AI search" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;p&gt;&lt;span style="font-weight: bold;"&gt;Claude and I had a 12,000 word back and forth on how to use Cloudflare and GA4 data to answer a hard question.&lt;/span&gt; If you were a SaaS marketer what would you do with all that context to make better answers?&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/saas-marketers-promote-your-claude-or-codex-connectors" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Wil%20Blog%20OG%20(6)-png.png" alt="SaaS Marketers - Promote your Claude &amp;amp; Codex connectors &amp;amp; win in AI search" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;p&gt;&lt;span style="font-weight: bold;"&gt;Claude and I had a 12,000 word back and forth on how to use Cloudflare and GA4 data to answer a hard question.&lt;/span&gt; If you were a SaaS marketer what would you do with all that context to make better answers?&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fsaas-marketers-promote-your-claude-or-codex-connectors&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>author-wil-reynolds</category>
      <category>AI</category>
      <pubDate>Mon, 28 Sep 2026 20:30:29 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/saas-marketers-promote-your-claude-or-codex-connectors</guid>
      <dc:date>2026-09-28T20:30:29Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>AI Citations Aren't Your Top KPI: ChatGPT-5.6 Luna Runs More Targeted Searches</title>
      <link>https://www.seerinteractive.com/insights/gpt-5.5-vs-gpt-5.6-luna</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/gpt-5.5-vs-gpt-5.6-luna" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/ChatGPT-5.6%20Luna%20Blog.png" alt="AI Citations Aren't Your Top KPI: ChatGPT-5.6 Luna Runs More Targeted Searches" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h3&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h3&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;We ran a direct comparison of 346 matched prompts to see whether ChatGPT-5.6 Luna behaved differently from GPT-5.5&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;GPT-5.6 Luna used web search more often but ran fewer queries, resulting in little movement on the number of fan-out queries compared to GPT-5.5&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Domain-target searching increased drastically (more than 3x), while consulted sources and inline citations dropped&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Our take: Luna is getting better at finding information and needs fewer resources, so brands should pay more attention to the full retrieval process to measure AI search visibility&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;How Do ChatGPT-5.5 and GPT-5.6 Differ?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;We’ve researched ChatGPT-5.5 extensively, including a &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/insights/the-proof-chatgpt-5.5s-fanout-patterns-reveal-the-importance-of-brand"&gt;&lt;u&gt;&lt;span&gt;deep dive into fan-out patterns&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt;. Now that GPT-5.6 Luna is the default model as of July 9, 2026, it's the model people get when interacting with ChatGPT and the model brands will be reporting on through AI visibility trackers.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;With that in mind, we wanted to know how the two models compared. Specifically: &lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;span&gt;How did the retrieval process change across domains and citations? &lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Does GPT-5.6 Luna search similarly, or were there notable differences?&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Did fan-out query behavior shift or stay the same? &lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;We compared GPT-5.5 and GPT-5.6 Luna across 346 matched prompts from six industries. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Spoiler alert: &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;At first glance, the two models looked very similar. Across the 346 matched prompts, they ran almost exactly the same number of fan-out queries. The difference is how targeted Luna's searches are. It used the site: operator in 76% of its fan-out queries, up from 22% with GPT-5.5.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That leaves far fewer open searches for everyone else. So if Luna isn't already searching your site directly, your barrier to showing up is higher than it used to be. Check out our top findings below.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;1. GPT-5.6 Luna Uses Web Search More Often, But Runs Fewer Queries Per Search&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;GPT-5.6 Luna turned to web search more often than GPT-5.5. Across the six industries, Luna searched on 85.2% of calls, up from 80.8%. But each time it searched, it ran fewer fan-out queries: 3.42 per search, down from 3.63.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Those two changes almost exactly cancel each other out. Luna searched more often but ran fewer queries each time, so the total number of fan-out queries barely moved. GPT-5.5 ran 733 and Luna ran 729, a difference of just four.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;We thought there would be more fan-out queries after the model update, especially given the increase we saw when GPT-5.5 was released. Instead, the total stayed the same, and the change showed up in how Luna searched.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;Our hypothesis is that Luna needs fewer queries per search because its queries are more targeted.&lt;/span&gt; When a fan-out query goes straight to a specific website (more on that in the next finding), one query can do the job that used to take several broader ones. So the two changes balance out in volume, but not in effect. More prompts now send ChatGPT to the web, and each search covers less ground.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;2. Domain-Targeted Searching More Than Tripled, Especially on Non-Brand Prompts&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;GPT-5.5 had already begun using the site: operator (a search command that limits results to a single website) in fan-out queries to target specific brand sites for answers and, in our opinion, prevent low-quality responses. GPT-5.6 Luna elevated this retrieval method even further.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;GPT-5.5 conducted site: searches in about 22% of its fan-out queries, while Luna did so in 76%.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;On branded prompts, third-party sites (publishers, review sites, research orgs, etc.) appeared at roughly the same rate with both models: 26.6% with GPT-5.5 and 24.4% with Luna. So even when a user asks about your brand directly, ChatGPT still looks beyond your website. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That means &lt;/span&gt;&lt;strong&gt;&lt;span&gt;the perception of your brand outside of your website matters just as much&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; as the narrative you share on sources you control. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The actual jump came from prompts that didn't mention a brand. Non-brand prompts saw a rise in site: searches, from 20.6% in GPT-5.5 to 77.3% in GPT-5.6. That means Luna was the one to decide on a domain to search even when the user didn’t specify a brand or website.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Since ChatGPT is getting more targeted with its retrieval, the overall number of consulted sources (the sources ChatGPT pulls in while searching, before it decides which ones to cite) fell 43%, from 6,134 with GPT-5.5 to 3,481 with Luna.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;This presents a unique opportunity for brands launching AI search programs. If GPT-5.6 Luna is already targeting your site with site: searches, make sure the information it's looking for is easy for both people and AI to find. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If your brand isn't one of the sites Luna searches directly, you may see fewer citations now that it's the default model. Open searches, the ones any site could appear in, fell from 575 to 175, so getting results will likely take more work.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Look at every stage of retrieval, from fan-out queries to consulted sources to citations, to see how and where ChatGPT finds information about your brand.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;3. Inline Citations Fell 13%&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;We know Luna searched in a more targeted way and consulted fewer sources. In the final answers, it also gave fewer inline citations (the sources linked in the answer itself).&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Across all 346 matched prompts, GPT-5.5 produced 2,054 inline citations, while Luna produced 1,790, a decrease of 12.9%. Searching more often didn't translate into more citations.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Our take is that Luna has gotten better at knowing where to find information, so it needs fewer sources to back up its answers.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;This matters for any AI search program that uses citations as a KPI or as the success metric for an experiment.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; If your citations dropped after Luna became the default, that may have nothing to do with your strategy or the tests you have running. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Think of it like a Google algorithm update: the platform changed, and performance shifted with it. Mark the switch to Luna in your reporting so a platform-wide drop doesn't get read as a failed experiment.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;So What Should You Do Now?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;How can you apply these findings to your AI search strategy? Here are my recommendations:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Look further down the retrieval process when you’re measuring AI search visibility.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Most AI visibility trackers only report citations, the bottom line of the retrieval process. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The bigger benefit comes from the layers behind them: the fan-out queries ChatGPT runs, the domains it targets with site: searches, and the sources it consults before deciding what to cite. That's where you'll see why your citations moved, not just that they did.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Pay closer attention to the connections the models appear to make between brands and topics. &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Luna often decides which sites to search before it retrieves anything, so being recognized as a relevant source for a topic matters before any page gets retrieved.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Run a base set of nonbrand prompts you’ve been monitoring to capture the fanout queries and sort the site: searches by topic. Which topics does ChatGPT search your domain for, and which does it search your competitors' domains for instead? &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Compare across models or over time, and you'll see where you're gaining ground and where a competitor owns the association. The topics your competitors own are where you should focus next: both in your own content, and in your coverage on publishers, review sites, and research organizations.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Ready to level up your AI search strategy but don’t know where to start? The Seer team has extensively researched what works and what’s a short-term trend, and we can recommend where to put your resources based on your industry, budget, and team size. &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/contact"&gt;&lt;u&gt;&lt;span&gt;Chat with us&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; to learn more.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/gpt-5.5-vs-gpt-5.6-luna" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/ChatGPT-5.6%20Luna%20Blog.png" alt="AI Citations Aren't Your Top KPI: ChatGPT-5.6 Luna Runs More Targeted Searches" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h3&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h3&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;We ran a direct comparison of 346 matched prompts to see whether ChatGPT-5.6 Luna behaved differently from GPT-5.5&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;GPT-5.6 Luna used web search more often but ran fewer queries, resulting in little movement on the number of fan-out queries compared to GPT-5.5&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Domain-target searching increased drastically (more than 3x), while consulted sources and inline citations dropped&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Our take: Luna is getting better at finding information and needs fewer resources, so brands should pay more attention to the full retrieval process to measure AI search visibility&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;How Do ChatGPT-5.5 and GPT-5.6 Differ?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;We’ve researched ChatGPT-5.5 extensively, including a &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/insights/the-proof-chatgpt-5.5s-fanout-patterns-reveal-the-importance-of-brand"&gt;&lt;u&gt;&lt;span&gt;deep dive into fan-out patterns&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt;. Now that GPT-5.6 Luna is the default model as of July 9, 2026, it's the model people get when interacting with ChatGPT and the model brands will be reporting on through AI visibility trackers.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;With that in mind, we wanted to know how the two models compared. Specifically: &lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;span&gt;How did the retrieval process change across domains and citations? &lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Does GPT-5.6 Luna search similarly, or were there notable differences?&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Did fan-out query behavior shift or stay the same? &lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;We compared GPT-5.5 and GPT-5.6 Luna across 346 matched prompts from six industries. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Spoiler alert: &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;At first glance, the two models looked very similar. Across the 346 matched prompts, they ran almost exactly the same number of fan-out queries. The difference is how targeted Luna's searches are. It used the site: operator in 76% of its fan-out queries, up from 22% with GPT-5.5.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That leaves far fewer open searches for everyone else. So if Luna isn't already searching your site directly, your barrier to showing up is higher than it used to be. Check out our top findings below.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;1. GPT-5.6 Luna Uses Web Search More Often, But Runs Fewer Queries Per Search&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;GPT-5.6 Luna turned to web search more often than GPT-5.5. Across the six industries, Luna searched on 85.2% of calls, up from 80.8%. But each time it searched, it ran fewer fan-out queries: 3.42 per search, down from 3.63.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Those two changes almost exactly cancel each other out. Luna searched more often but ran fewer queries each time, so the total number of fan-out queries barely moved. GPT-5.5 ran 733 and Luna ran 729, a difference of just four.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;We thought there would be more fan-out queries after the model update, especially given the increase we saw when GPT-5.5 was released. Instead, the total stayed the same, and the change showed up in how Luna searched.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;Our hypothesis is that Luna needs fewer queries per search because its queries are more targeted.&lt;/span&gt; When a fan-out query goes straight to a specific website (more on that in the next finding), one query can do the job that used to take several broader ones. So the two changes balance out in volume, but not in effect. More prompts now send ChatGPT to the web, and each search covers less ground.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;2. Domain-Targeted Searching More Than Tripled, Especially on Non-Brand Prompts&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;GPT-5.5 had already begun using the site: operator (a search command that limits results to a single website) in fan-out queries to target specific brand sites for answers and, in our opinion, prevent low-quality responses. GPT-5.6 Luna elevated this retrieval method even further.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;GPT-5.5 conducted site: searches in about 22% of its fan-out queries, while Luna did so in 76%.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;On branded prompts, third-party sites (publishers, review sites, research orgs, etc.) appeared at roughly the same rate with both models: 26.6% with GPT-5.5 and 24.4% with Luna. So even when a user asks about your brand directly, ChatGPT still looks beyond your website. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That means &lt;/span&gt;&lt;strong&gt;&lt;span&gt;the perception of your brand outside of your website matters just as much&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; as the narrative you share on sources you control. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The actual jump came from prompts that didn't mention a brand. Non-brand prompts saw a rise in site: searches, from 20.6% in GPT-5.5 to 77.3% in GPT-5.6. That means Luna was the one to decide on a domain to search even when the user didn’t specify a brand or website.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Since ChatGPT is getting more targeted with its retrieval, the overall number of consulted sources (the sources ChatGPT pulls in while searching, before it decides which ones to cite) fell 43%, from 6,134 with GPT-5.5 to 3,481 with Luna.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;This presents a unique opportunity for brands launching AI search programs. If GPT-5.6 Luna is already targeting your site with site: searches, make sure the information it's looking for is easy for both people and AI to find. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If your brand isn't one of the sites Luna searches directly, you may see fewer citations now that it's the default model. Open searches, the ones any site could appear in, fell from 575 to 175, so getting results will likely take more work.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Look at every stage of retrieval, from fan-out queries to consulted sources to citations, to see how and where ChatGPT finds information about your brand.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;3. Inline Citations Fell 13%&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;We know Luna searched in a more targeted way and consulted fewer sources. In the final answers, it also gave fewer inline citations (the sources linked in the answer itself).&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Across all 346 matched prompts, GPT-5.5 produced 2,054 inline citations, while Luna produced 1,790, a decrease of 12.9%. Searching more often didn't translate into more citations.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Our take is that Luna has gotten better at knowing where to find information, so it needs fewer sources to back up its answers.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;This matters for any AI search program that uses citations as a KPI or as the success metric for an experiment.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; If your citations dropped after Luna became the default, that may have nothing to do with your strategy or the tests you have running. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Think of it like a Google algorithm update: the platform changed, and performance shifted with it. Mark the switch to Luna in your reporting so a platform-wide drop doesn't get read as a failed experiment.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;So What Should You Do Now?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;How can you apply these findings to your AI search strategy? Here are my recommendations:&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Look further down the retrieval process when you’re measuring AI search visibility.&lt;/span&gt;&lt;/strong&gt;&lt;span&gt; Most AI visibility trackers only report citations, the bottom line of the retrieval process. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The bigger benefit comes from the layers behind them: the fan-out queries ChatGPT runs, the domains it targets with site: searches, and the sources it consults before deciding what to cite. That's where you'll see why your citations moved, not just that they did.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;strong&gt;&lt;span&gt;Pay closer attention to the connections the models appear to make between brands and topics. &lt;/span&gt;&lt;/strong&gt;&lt;span&gt;Luna often decides which sites to search before it retrieves anything, so being recognized as a relevant source for a topic matters before any page gets retrieved.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Run a base set of nonbrand prompts you’ve been monitoring to capture the fanout queries and sort the site: searches by topic. Which topics does ChatGPT search your domain for, and which does it search your competitors' domains for instead? &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Compare across models or over time, and you'll see where you're gaining ground and where a competitor owns the association. The topics your competitors own are where you should focus next: both in your own content, and in your coverage on publishers, review sites, and research organizations.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Ready to level up your AI search strategy but don’t know where to start? The Seer team has extensively researched what works and what’s a short-term trend, and we can recommend where to put your resources based on your industry, budget, and team size. &lt;/span&gt;&lt;a href="https://www.seerinteractive.com/contact"&gt;&lt;u&gt;&lt;span&gt;Chat with us&lt;/span&gt;&lt;/u&gt;&lt;/a&gt;&lt;span&gt; to learn more.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&amp;nbsp;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fgpt-5.5-vs-gpt-5.6-luna&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>AI</category>
      <category>Author: Bryan Gunawan</category>
      <pubDate>Thu, 24 Sep 2026 19:36:11 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/gpt-5.5-vs-gpt-5.6-luna</guid>
      <dc:date>2026-09-24T19:36:11Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>Our Most Valuable Blogs of 2026: According to Our Readers</title>
      <link>https://www.seerinteractive.com/insights/our-most-valuable-blogs-of-2026-according-to-our-readers</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/our-most-valuable-blogs-of-2026-according-to-our-readers" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Blog%20Update%20%231%20(1).png" alt="Our Most Valuable Blogs of 2026: According to Our Readers" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;p&gt;&lt;span style="font-weight: bold;"&gt;What's my favorite horror movie?&lt;/span&gt; When I patted myself on the back for setting up an ~autonomous workflow~ that kept our newsletter subscriber hygiene clean by automatically unsubscribing anyone who hasn't opened any of our emails in the past 12 months.&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/our-most-valuable-blogs-of-2026-according-to-our-readers" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Blog%20Update%20%231%20(1).png" alt="Our Most Valuable Blogs of 2026: According to Our Readers" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;p&gt;&lt;span style="font-weight: bold;"&gt;What's my favorite horror movie?&lt;/span&gt; When I patted myself on the back for setting up an ~autonomous workflow~ that kept our newsletter subscriber hygiene clean by automatically unsubscribing anyone who hasn't opened any of our emails in the past 12 months.&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Four-most-valuable-blogs-of-2026-according-to-our-readers&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>author-hannah-cooley</category>
      <category>AI</category>
      <pubDate>Tue, 22 Sep 2026 21:00:18 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/our-most-valuable-blogs-of-2026-according-to-our-readers</guid>
      <dc:date>2026-09-22T21:00:18Z</dc:date>
      <dc:creator>Hannah Cooley</dc:creator>
    </item>
    <item>
      <title>Are You Struggling with Market Pressure or Operational Readiness? Use Our 2x2 Framework to Find Out</title>
      <link>https://www.seerinteractive.com/insights/maturity-model-framework</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/maturity-model-framework" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Market%20vs%20operational%20blog.png" alt="Are You Struggling with Market Pressure or Operational Readiness? Use Our 2x2 Framework to Find Out" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li&gt;If your performance numbers are down, your single-axis maturity model might be telling you to solve for the wrong problem&lt;/li&gt; 
 &lt;li&gt;A true framework should factor in two variables: what the market is doing to you (market pressure), and what you're equipped to do about it (operational readiness)&lt;/li&gt; 
 &lt;li&gt;Use our 2x2 framework (and our handy worksheet) that considers both variables to more accurately evaluate where you fit&lt;/li&gt; 
 &lt;li&gt;Two Seer clients thought they had performance problems but used our framework to pinpoint other underlying issues&lt;/li&gt; 
 &lt;li&gt;Learn three quick things you can do this week to help identify the underlying issues that could be impacting your numbers&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;Is Your Maturity Score Focusing on the Wrong Problem?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;I've worked with dozens of marketing teams, and every one of them has had the same bad week. Numbers dip, costs climb, and someone asks what happened.&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/maturity-model-framework" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Market%20vs%20operational%20blog.png" alt="Are You Struggling with Market Pressure or Operational Readiness? Use Our 2x2 Framework to Find Out" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li&gt;If your performance numbers are down, your single-axis maturity model might be telling you to solve for the wrong problem&lt;/li&gt; 
 &lt;li&gt;A true framework should factor in two variables: what the market is doing to you (market pressure), and what you're equipped to do about it (operational readiness)&lt;/li&gt; 
 &lt;li&gt;Use our 2x2 framework (and our handy worksheet) that considers both variables to more accurately evaluate where you fit&lt;/li&gt; 
 &lt;li&gt;Two Seer clients thought they had performance problems but used our framework to pinpoint other underlying issues&lt;/li&gt; 
 &lt;li&gt;Learn three quick things you can do this week to help identify the underlying issues that could be impacting your numbers&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;Is Your Maturity Score Focusing on the Wrong Problem?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;I've worked with dozens of marketing teams, and every one of them has had the same bad week. Numbers dip, costs climb, and someone asks what happened.&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fmaturity-model-framework&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Paid Media</category>
      <category>Author: Carlene Lombo</category>
      <pubDate>Thu, 10 Sep 2026 16:03:54 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/maturity-model-framework</guid>
      <dc:date>2026-09-10T16:03:54Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>Google Is Not Going to Stop Ranking Your Claude-Assisted Content</title>
      <link>https://www.seerinteractive.com/insights/anthropic-watermark-update</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/anthropic-watermark-update" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Claude%20Watermark%20Update%20(1).png" alt="Google Is Not Going to Stop Ranking Your Claude-Assisted Content" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2 style="line-height: 1.2;"&gt;&lt;a&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/a&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li&gt;Claude’s watermark only flags that Claude was “likely involved,” so it can’t tell a fully AI-drafted piece from a human draft Claude lightly edited&lt;/li&gt; 
 &lt;li&gt;It’s an EU compliance requirement (Code of Practice on Transparency of AI-Generated Content). This is not a new Google ranking signal.&lt;/li&gt; 
 &lt;li&gt;Google has run its own version (SynthID-text)on Gemini for two years and never used it to rank anything&lt;/li&gt; 
 &lt;li&gt;The real traffic risk is thin, templated content and not the watermark. Sites that scaled content off a copy-the-template approach saw more than half those lose 30%+ of peak organic traffic&lt;/li&gt; 
 &lt;li&gt;What still matters: unique inputs and a human accountable for the final piece&lt;/li&gt; 
&lt;/ul&gt;  
&lt;p&gt;&lt;span&gt;Anthropic started watermarking Claude's text output. &lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/anthropic-watermark-update" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Claude%20Watermark%20Update%20(1).png" alt="Google Is Not Going to Stop Ranking Your Claude-Assisted Content" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2 style="line-height: 1.2;"&gt;&lt;a&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/a&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li&gt;Claude’s watermark only flags that Claude was “likely involved,” so it can’t tell a fully AI-drafted piece from a human draft Claude lightly edited&lt;/li&gt; 
 &lt;li&gt;It’s an EU compliance requirement (Code of Practice on Transparency of AI-Generated Content). This is not a new Google ranking signal.&lt;/li&gt; 
 &lt;li&gt;Google has run its own version (SynthID-text)on Gemini for two years and never used it to rank anything&lt;/li&gt; 
 &lt;li&gt;The real traffic risk is thin, templated content and not the watermark. Sites that scaled content off a copy-the-template approach saw more than half those lose 30%+ of peak organic traffic&lt;/li&gt; 
 &lt;li&gt;What still matters: unique inputs and a human accountable for the final piece&lt;/li&gt; 
&lt;/ul&gt;  
&lt;p&gt;&lt;span&gt;Anthropic started watermarking Claude's text output. &lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fanthropic-watermark-update&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Author: Alisa Scharf</category>
      <category>AI</category>
      <category>LLM</category>
      <pubDate>Fri, 04 Sep 2026 18:24:31 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/anthropic-watermark-update</guid>
      <dc:date>2026-09-04T18:24:31Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>AI Can Do More for Project Management Than Note-Taking: 4 Ways to Level-Up Your Usage</title>
      <link>https://www.seerinteractive.com/insights/how-project-managers-use-ai</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/how-project-managers-use-ai" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/AI%20for%20Project%20Mgmt%20(1).png" alt="AI Can Do More for Project Management Than Note-Taking: 4 Ways to Level-Up Your Usage" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h3&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h3&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;AI is more than just a summary or synthesis tool for project managers, and the best PMs will use it to become a better strategic partner&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;PMs can use AI to spot ongoing friction and pinpoint the root cause, or&lt;/span&gt;&lt;span&gt;&amp;nbsp;look for and resolve broader patterns across your client accounts&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Use AI as a sounding board to find the gap in your idea before bringing it to the client&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Leverage internal expertise and documentation to build client-ready deliverables with minimal manpower&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;AI can't interpret human behavior or responses, and that's where PMs can shine when they offload busywork to AI&amp;nbsp;&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;Are PMs Extracting the Maximum Value from AI?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;By now, most project managers are using AI to pull action items from a call, draft a status update, or help you write an SOP.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;But if the tactical execution part of the job has gotten cheaper and faster thanks to AI, what's left for a PM to be good at? &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I think the answer is the thing most of us got into this work hoping to do in the first place: being a genuine strategic partner to the client, not just the person who keeps the trains running on time.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here are 4 ways I’ve started using AI to level-up my client strategies and build stronger client relationships.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;1. Catching the Pattern You'd Never Notice Call by Call&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Our CEO, Wil Reynolds, says something that’s stuck with me: use AI for the manual tasks your brain power far exceeds, and spend that brain power on strategy and moving the needle for the client. I didn’t fully get what that meant until I started actually doing it. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Now it's part of my process. After almost every client call, I ask AI one question based on the transcript:&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #fdfaf2; border-left: 3px solid #f5cc52; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;span style="font-weight: bold;"&gt;THE PROMPT&lt;/span&gt;&lt;br&gt;&lt;em&gt;How can we be a better strategic partner here, based on what the client said?&lt;/em&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Just asking AI to dig for subtext instead of a summary isn't a differentiator anymore. Plenty of PMs have figured out that part. What's more impactful is what you do with whatever AI surfaces.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;In practice:&lt;/span&gt; With that prompt, AI helped me understand the root of recurring campaign delays for one client: they didn’t know where our campaigns stood, and approval requests were scattered across Slack threads with no clear queue. They weren’t stalling on purpose, they just had no single place to understand what was needed from them.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;After AI picked up on the pattern, I worked on a solution and built a live campaign tracker. Now, the client had:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;span&gt;A scheduled weekly update via Slack on where every campaign stood&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;A dedicated approvals-only Slack channel for our team to share exactly what we needed from them&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Updates formatted so the client could scan, review, and approve in minutes &lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;The client stopped being an accidental bottleneck and started closing out approvals in minutes, and our team stopped losing time chasing sign-off across multiple channels. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;A single call would never have told me that on its own. It took AI flagging the same friction call after call before the real fix became obvious.&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #fdfaf2; border-left: 3px solid #f5cc52; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;strong&gt;TRY THIS&lt;/strong&gt;&lt;br&gt;After your next few client calls, run each transcript through the same question. Focus not on what was said, but on how you could be a better strategic partner. Don't act on the first answer. Watch for the same friction showing up more than once before you build anything around it.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;2. Seeing Across Accounts, Not Just Within One&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Most PMs manage more than one relationship, and the thing AI is uniquely good at that a busy human isn't, is holding all of them in view at once. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;So I started asking a version of the same question I use on individual calls, but pointed at my whole book of business: is there anything showing up across more than one of my accounts that I'm not connecting?&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;In practice:&lt;/span&gt; AI surfaced that GEO test implementation was stalling for a similar reason across a few accounts that, on paper, had nothing in common. It wasn't a technical blocker. It was resourcing; client teams didn't have the bandwidth to execute anything complex, and that was only visible once I stopped looking at each account in isolation.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Seeing that bigger issue led me to change how I approached a GEO project for one client in particular. Implementation had already stalled. Tests weren't moving because the work assumed more bandwidth than the client's team actually had.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Instead of continuing to push on the same complex tests, I raised the resourcing question directly. The client confirmed their team's bandwidth was limited right now, which told me exactly what needed to change: I had my internal team shift focus to easy, low-lift test ideas the client could realistically execute. That shift is what got the project back on track.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That pattern only showed up because AI helped me look beyond my individual accounts when I didn’t have the bandwidth to hold everything in view all at once.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #c7eceb; border-left: 3px solid #54dedb; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;strong&gt;TRY THIS&lt;/strong&gt;&lt;br&gt;If you manage more than one account, ask AI a version of this before your next resourcing or roadmap conversation: is anything showing up across more than one of my accounts that I've been treating as separate? Do it before something stalls, not after.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;3. Using AI to Pressure-Test Yourself Before the Client Does&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;AI has a default instinct to agree with you. It wants to be helpful, which in practice often means it wants to validate whatever you just said. Left unchecked, that's the opposite of useful when you're trying to stress-test a strategy.&amp;nbsp;&lt;br&gt;&lt;br&gt;So I set a standing rule for how mine should behave:&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Don't agree by default&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Find the weakest point in what I've said before you affirm any of it&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Tell me what someone who disagrees would say&amp;nbsp;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;When I bring AI a strategy or ask for a read on a situation, its first move is to argue with me, not validate me. That changes the conversation completely. Instead of a tool that confirms what I already think, it's one that makes me defend my idea first.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;In practice: &lt;/span&gt;A new contact on one of my accounts wanted the relationship with our team to feel more collaborative. She was interested in less status reporting and more working through things together.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;My first instinct was straightforward: change our current meeting format. Turn the recurring status call into more of a strategic discussion.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I ran that call's transcript through AI afterward, and it didn't just validate the plan; it found the gap in it. Buried in the same conversation, my own team had voiced frustration about how campaigns were getting launched: almost no input from us on strategy, and asset timelines so tight there was no room to do more than execute whatever landed.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;AI's read: changing the meeting format was a reasonable start, but it wouldn't touch the actual problem. If we wanted to be real collaborators, the meetings needed space to shape campaign strategy together, talk through how assets should be positioned by platform, and get ahead of sizing and formatting before assets showed up with days to spare.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That's the version we actually built: a new SOP centered on recurring strategy sessions, not a lighter-touch status call. The relationship is strong enough now that the client's asking about expanding our scope to additional divisions because we became the partner she wanted.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;My instinct wasn't wrong. But the fix needed to go further, and I wouldn't have caught that on a single read of my own notes.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #ffd5bc; border-left: 3px solid #ff9033; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;strong&gt;TRY THIS&lt;/strong&gt;&lt;br&gt;Before you bring your next recommendation to a client, run it past AI with the same standing instruction: don't agree, find the weakest point, tell me what someone who disagrees would say. If it doesn't push back at all, that's worth noticing as it usually means the instruction didn't take, not that your plan is airtight.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;4. Spearheading the Work Instead of Coordinating It&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;The clearest example of what this actually changes for a PM happened when a client asked for a full deck to present in a CMO meeting that had landed on their calendar with about 24 hours' notice. My team was stretched thin elsewhere, and we didn’t have the bandwidth to pull together a normal cross-functional build.&lt;br&gt;&lt;br&gt;&lt;span style="font-weight: bold;"&gt;In practice: &lt;/span&gt;I built the deck myself by taking our existing analysis decks, call transcripts, and Slack threads for the account, feeding all of it into AI, and specifying the story I wanted the deck to tell. Not “summarize this,” but the actual narrative: here’s what matters to this CMO, here’s the through-line I want a stakeholder walking away with.&amp;nbsp;&lt;br&gt;&lt;br&gt;The point wasn’t about AI being fast. It was that I had the full context on the account: the relationship history, the priorities, and the political read on the room. AI closed the gap between having that context and having a polished deliverable, without needing to pull in three other people to get there.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;It used to take a full cross-functional build to turn scattered account knowledge into a client-ready strategic narrative. Increasingly, a PM who already holds that knowledge can spearhead it, carrying the team’s work into the room rather than convening the room to package it.&amp;nbsp;&lt;br&gt;To be clear: AI wasn’t replacing our specialists, but rather leveraging the work our specialists had already done to inform the output.&amp;nbsp;&lt;br&gt;&lt;br&gt;Those analysis decks, the account knowledge, the read on the CMO all came from the team’s expertise. Because I already had the information, I was able to turn it into a polished, client-ready narrative on a 24-hour turnaround.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The layer this shrinks is synthesis and assembly, not subject-matter expertise. And that’s the opportunity: a lot of cross-functional time gets spent packaging what someone already knows. If the PM can spearhead that layer, our specialists get more of their time back for the deep work only they can do.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #e5b8fc; border-left: 3px solid #ba38ff; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;strong&gt;TRY THIS&lt;/strong&gt;&lt;br&gt;Next time you’re under deadline with everything you need already scattered across decks, transcripts, and Slack, don’t ask AI to summarize it. Tell it the actual narrative you want a stakeholder to walk away with, and let it build toward that. See how far you get solo before asking for backup.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;What AI Still Can’t Do: Interpret Human Behavior or Responses&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;None of this makes the PM optional.&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;For example, go back to that campaign tracker and approvals channel from Pattern 1: AI could see that requests were stuck and tell me exactly where. But it couldn't tell me whether raising that with the client directly would land as helpful or harmful.&lt;/p&gt; 
&lt;p&gt;That read came from a dozen calls with the client, understanding: how he takes feedback, what the relationship could hold at that point, whether this was a conversation for a call or a quiet Slack message. None of that lives anywhere AI can find it.&lt;/p&gt; 
&lt;p&gt;AI is good at telling you something's there. It has no way of telling you how it'll land with this specific person, right now. That part is still entirely yours.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;Where This Leaves the Project Manager Role&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;For years, the job rewarded whoever could juggle the most balls. That was never the point.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The PMs who get the most out of AI are using the space that creates to actually think: about the account, about the client's business, about what they're not being told directly.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here's where I think this actually takes the role over the next year:&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Less time spent proving the work got done&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Reading what a client isn't saying&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Knowing when a relationship needs a different kind of attention than a status update&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Building the judgment calls that no prompt can make for you&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;The PMs building habits like the ones above will spend more time doing the work that made them want this job in the first place, while the busywork keeps quietly shrinking around them.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That's the opening in front of every PM right now: less of your day spent holding it together, more of it spent actually thinking, and finally the room to become the strategic partner you always wanted to be.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you're looking for a strategic partner to level-up your marketing activities, Seer can help. &lt;a href="https://www.seerinteractive.com/contact"&gt;Chat with our team&lt;/a&gt; to learn what we can achieve together.&amp;nbsp;&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/how-project-managers-use-ai" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/AI%20for%20Project%20Mgmt%20(1).png" alt="AI Can Do More for Project Management Than Note-Taking: 4 Ways to Level-Up Your Usage" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h3&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h3&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;AI is more than just a summary or synthesis tool for project managers, and the best PMs will use it to become a better strategic partner&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;PMs can use AI to spot ongoing friction and pinpoint the root cause, or&lt;/span&gt;&lt;span&gt;&amp;nbsp;look for and resolve broader patterns across your client accounts&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Use AI as a sounding board to find the gap in your idea before bringing it to the client&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Leverage internal expertise and documentation to build client-ready deliverables with minimal manpower&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;AI can't interpret human behavior or responses, and that's where PMs can shine when they offload busywork to AI&amp;nbsp;&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;Are PMs Extracting the Maximum Value from AI?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;By now, most project managers are using AI to pull action items from a call, draft a status update, or help you write an SOP.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;But if the tactical execution part of the job has gotten cheaper and faster thanks to AI, what's left for a PM to be good at? &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I think the answer is the thing most of us got into this work hoping to do in the first place: being a genuine strategic partner to the client, not just the person who keeps the trains running on time.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here are 4 ways I’ve started using AI to level-up my client strategies and build stronger client relationships.&lt;/span&gt;&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;1. Catching the Pattern You'd Never Notice Call by Call&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Our CEO, Wil Reynolds, says something that’s stuck with me: use AI for the manual tasks your brain power far exceeds, and spend that brain power on strategy and moving the needle for the client. I didn’t fully get what that meant until I started actually doing it. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Now it's part of my process. After almost every client call, I ask AI one question based on the transcript:&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #fdfaf2; border-left: 3px solid #f5cc52; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;span style="font-weight: bold;"&gt;THE PROMPT&lt;/span&gt;&lt;br&gt;&lt;em&gt;How can we be a better strategic partner here, based on what the client said?&lt;/em&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Just asking AI to dig for subtext instead of a summary isn't a differentiator anymore. Plenty of PMs have figured out that part. What's more impactful is what you do with whatever AI surfaces.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;In practice:&lt;/span&gt; With that prompt, AI helped me understand the root of recurring campaign delays for one client: they didn’t know where our campaigns stood, and approval requests were scattered across Slack threads with no clear queue. They weren’t stalling on purpose, they just had no single place to understand what was needed from them.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;After AI picked up on the pattern, I worked on a solution and built a live campaign tracker. Now, the client had:&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;span&gt;A scheduled weekly update via Slack on where every campaign stood&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;A dedicated approvals-only Slack channel for our team to share exactly what we needed from them&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Updates formatted so the client could scan, review, and approve in minutes &lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;The client stopped being an accidental bottleneck and started closing out approvals in minutes, and our team stopped losing time chasing sign-off across multiple channels. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;A single call would never have told me that on its own. It took AI flagging the same friction call after call before the real fix became obvious.&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #fdfaf2; border-left: 3px solid #f5cc52; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;strong&gt;TRY THIS&lt;/strong&gt;&lt;br&gt;After your next few client calls, run each transcript through the same question. Focus not on what was said, but on how you could be a better strategic partner. Don't act on the first answer. Watch for the same friction showing up more than once before you build anything around it.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;2. Seeing Across Accounts, Not Just Within One&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;Most PMs manage more than one relationship, and the thing AI is uniquely good at that a busy human isn't, is holding all of them in view at once. &lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;So I started asking a version of the same question I use on individual calls, but pointed at my whole book of business: is there anything showing up across more than one of my accounts that I'm not connecting?&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;In practice:&lt;/span&gt; AI surfaced that GEO test implementation was stalling for a similar reason across a few accounts that, on paper, had nothing in common. It wasn't a technical blocker. It was resourcing; client teams didn't have the bandwidth to execute anything complex, and that was only visible once I stopped looking at each account in isolation.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Seeing that bigger issue led me to change how I approached a GEO project for one client in particular. Implementation had already stalled. Tests weren't moving because the work assumed more bandwidth than the client's team actually had.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Instead of continuing to push on the same complex tests, I raised the resourcing question directly. The client confirmed their team's bandwidth was limited right now, which told me exactly what needed to change: I had my internal team shift focus to easy, low-lift test ideas the client could realistically execute. That shift is what got the project back on track.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That pattern only showed up because AI helped me look beyond my individual accounts when I didn’t have the bandwidth to hold everything in view all at once.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #c7eceb; border-left: 3px solid #54dedb; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;strong&gt;TRY THIS&lt;/strong&gt;&lt;br&gt;If you manage more than one account, ask AI a version of this before your next resourcing or roadmap conversation: is anything showing up across more than one of my accounts that I've been treating as separate? Do it before something stalls, not after.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;3. Using AI to Pressure-Test Yourself Before the Client Does&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;AI has a default instinct to agree with you. It wants to be helpful, which in practice often means it wants to validate whatever you just said. Left unchecked, that's the opposite of useful when you're trying to stress-test a strategy.&amp;nbsp;&lt;br&gt;&lt;br&gt;So I set a standing rule for how mine should behave:&amp;nbsp;&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Don't agree by default&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Find the weakest point in what I've said before you affirm any of it&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Tell me what someone who disagrees would say&amp;nbsp;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;When I bring AI a strategy or ask for a read on a situation, its first move is to argue with me, not validate me. That changes the conversation completely. Instead of a tool that confirms what I already think, it's one that makes me defend my idea first.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;&lt;span style="font-weight: bold;"&gt;In practice: &lt;/span&gt;A new contact on one of my accounts wanted the relationship with our team to feel more collaborative. She was interested in less status reporting and more working through things together.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;My first instinct was straightforward: change our current meeting format. Turn the recurring status call into more of a strategic discussion.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;I ran that call's transcript through AI afterward, and it didn't just validate the plan; it found the gap in it. Buried in the same conversation, my own team had voiced frustration about how campaigns were getting launched: almost no input from us on strategy, and asset timelines so tight there was no room to do more than execute whatever landed.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;AI's read: changing the meeting format was a reasonable start, but it wouldn't touch the actual problem. If we wanted to be real collaborators, the meetings needed space to shape campaign strategy together, talk through how assets should be positioned by platform, and get ahead of sizing and formatting before assets showed up with days to spare.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That's the version we actually built: a new SOP centered on recurring strategy sessions, not a lighter-touch status call. The relationship is strong enough now that the client's asking about expanding our scope to additional divisions because we became the partner she wanted.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;My instinct wasn't wrong. But the fix needed to go further, and I wouldn't have caught that on a single read of my own notes.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #ffd5bc; border-left: 3px solid #ff9033; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;strong&gt;TRY THIS&lt;/strong&gt;&lt;br&gt;Before you bring your next recommendation to a client, run it past AI with the same standing instruction: don't agree, find the weakest point, tell me what someone who disagrees would say. If it doesn't push back at all, that's worth noticing as it usually means the instruction didn't take, not that your plan is airtight.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;4. Spearheading the Work Instead of Coordinating It&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;The clearest example of what this actually changes for a PM happened when a client asked for a full deck to present in a CMO meeting that had landed on their calendar with about 24 hours' notice. My team was stretched thin elsewhere, and we didn’t have the bandwidth to pull together a normal cross-functional build.&lt;br&gt;&lt;br&gt;&lt;span style="font-weight: bold;"&gt;In practice: &lt;/span&gt;I built the deck myself by taking our existing analysis decks, call transcripts, and Slack threads for the account, feeding all of it into AI, and specifying the story I wanted the deck to tell. Not “summarize this,” but the actual narrative: here’s what matters to this CMO, here’s the through-line I want a stakeholder walking away with.&amp;nbsp;&lt;br&gt;&lt;br&gt;The point wasn’t about AI being fast. It was that I had the full context on the account: the relationship history, the priorities, and the political read on the room. AI closed the gap between having that context and having a polished deliverable, without needing to pull in three other people to get there.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;It used to take a full cross-functional build to turn scattered account knowledge into a client-ready strategic narrative. Increasingly, a PM who already holds that knowledge can spearhead it, carrying the team’s work into the room rather than convening the room to package it.&amp;nbsp;&lt;br&gt;To be clear: AI wasn’t replacing our specialists, but rather leveraging the work our specialists had already done to inform the output.&amp;nbsp;&lt;br&gt;&lt;br&gt;Those analysis decks, the account knowledge, the read on the CMO all came from the team’s expertise. Because I already had the information, I was able to turn it into a polished, client-ready narrative on a 24-hour turnaround.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The layer this shrinks is synthesis and assembly, not subject-matter expertise. And that’s the opportunity: a lot of cross-functional time gets spent packaging what someone already knows. If the PM can spearhead that layer, our specialists get more of their time back for the deep work only they can do.&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p style="background-color: #e5b8fc; border-left: 3px solid #ba38ff; padding: 12px 16px; margin: 8px 0 16px 0;"&gt;&lt;strong&gt;TRY THIS&lt;/strong&gt;&lt;br&gt;Next time you’re under deadline with everything you need already scattered across decks, transcripts, and Slack, don’t ask AI to summarize it. Tell it the actual narrative you want a stakeholder to walk away with, and let it build toward that. See how far you get solo before asking for backup.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;What AI Still Can’t Do: Interpret Human Behavior or Responses&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;None of this makes the PM optional.&amp;nbsp;&lt;/p&gt; 
&lt;p&gt;For example, go back to that campaign tracker and approvals channel from Pattern 1: AI could see that requests were stuck and tell me exactly where. But it couldn't tell me whether raising that with the client directly would land as helpful or harmful.&lt;/p&gt; 
&lt;p&gt;That read came from a dozen calls with the client, understanding: how he takes feedback, what the relationship could hold at that point, whether this was a conversation for a call or a quiet Slack message. None of that lives anywhere AI can find it.&lt;/p&gt; 
&lt;p&gt;AI is good at telling you something's there. It has no way of telling you how it'll land with this specific person, right now. That part is still entirely yours.&lt;/p&gt; 
&lt;h2&gt;&lt;span&gt;Where This Leaves the Project Manager Role&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;For years, the job rewarded whoever could juggle the most balls. That was never the point.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;The PMs who get the most out of AI are using the space that creates to actually think: about the account, about the client's business, about what they're not being told directly.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Here's where I think this actually takes the role over the next year:&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;ul&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Less time spent proving the work got done&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Reading what a client isn't saying&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Knowing when a relationship needs a different kind of attention than a status update&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
 &lt;li&gt; &lt;p&gt;&lt;span&gt;Building the judgment calls that no prompt can make for you&amp;nbsp;&lt;br&gt;&lt;/span&gt;&lt;/p&gt; &lt;/li&gt; 
&lt;/ul&gt; 
&lt;p&gt;&lt;span&gt;The PMs building habits like the ones above will spend more time doing the work that made them want this job in the first place, while the busywork keeps quietly shrinking around them.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;That's the opening in front of every PM right now: less of your day spent holding it together, more of it spent actually thinking, and finally the room to become the strategic partner you always wanted to be.&lt;br&gt;&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;If you're looking for a strategic partner to level-up your marketing activities, Seer can help. &lt;a href="https://www.seerinteractive.com/contact"&gt;Chat with our team&lt;/a&gt; to learn what we can achieve together.&amp;nbsp;&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fhow-project-managers-use-ai&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Client Services</category>
      <category>Author: Jessica Simms</category>
      <pubDate>Wed, 02 Sep 2026 18:46:20 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/how-project-managers-use-ai</guid>
      <dc:date>2026-09-02T18:46:20Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>AI Visibility Is Lying to You: There’s No One-Size-Fits-All AI Search KPI</title>
      <link>https://www.seerinteractive.com/insights/ai-visibility-is-lying-to-you</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/ai-visibility-is-lying-to-you" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/GEO%20Metrics%20(2).png" alt="AI Visibility Is Lying to You: There’s No One-Size-Fits-All AI Search KPI" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;span&gt;Everyone measuring AI search is defaulting to the same KPIs, but there is no one-size-fits-all AI search KPI&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;New research shows 54% of AI search categories lack a clear category owner, and those categories represent 89% of all AI search demand&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;If someone else owns your category, you should stop chasing visibility and instead focus on accuracy: when you do show up, does the model describe you the way you actually want to be described?&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Accurately measuring AI search KPIs requires a brand canon, a living document of your brand’s claims, positioning, facts, and more&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;Your Visibility Score Can Be Perfect and Still Be Wrong&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;AI can mention a brand in every prompt you track, but that brand can still be losing.&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/ai-visibility-is-lying-to-you" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/GEO%20Metrics%20(2).png" alt="AI Visibility Is Lying to You: There’s No One-Size-Fits-All AI Search KPI" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;span&gt;Everyone measuring AI search is defaulting to the same KPIs, but there is no one-size-fits-all AI search KPI&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;New research shows 54% of AI search categories lack a clear category owner, and those categories represent 89% of all AI search demand&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;If someone else owns your category, you should stop chasing visibility and instead focus on accuracy: when you do show up, does the model describe you the way you actually want to be described?&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Accurately measuring AI search KPIs requires a brand canon, a living document of your brand’s claims, positioning, facts, and more&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;strong&gt;&lt;span&gt;Your Visibility Score Can Be Perfect and Still Be Wrong&lt;/span&gt;&lt;/strong&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;AI can mention a brand in every prompt you track, but that brand can still be losing.&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fai-visibility-is-lying-to-you&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>AI</category>
      <category>GEO</category>
      <category>Author: Sonny Vasquez</category>
      <pubDate>Mon, 31 Aug 2026 19:59:48 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/ai-visibility-is-lying-to-you</guid>
      <dc:date>2026-08-31T19:59:48Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>Study: AI Declined to Answer Prompts About Our Brand 31% of the Time (But Got Our Info Mostly Right When It Answered)</title>
      <link>https://www.seerinteractive.com/insights/ai-brand-accuracy-study</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/ai-brand-accuracy-study" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Brand%20Accuracy%20Study-1.png" alt="Study:&amp;nbsp;AI Declined to Answer Prompts About Our Brand 31% of the Time (But Got Our Info Mostly Right When It Answered)" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h3&gt;TL;DR&lt;/h3&gt; 
&lt;ul&gt; 
 &lt;li&gt;We realized AI platforms were relaying an inaccurate narrative about Seer Interactive, so we ran a full analysis to find out what else they were saying — and whether the info was accurate&lt;/li&gt; 
 &lt;li&gt;Of the 6 AI platforms we tracked, they averaged a 95.9% accuracy rating when they answered the prompt&lt;/li&gt; 
 &lt;li&gt;The AI platforms were more likely to decline to answer a prompt (31% on average) than to provide an incorrect answer about Seer (3%)&lt;/li&gt; 
 &lt;li&gt;AI frequently declined to answer prompts about comparisons (80%), but had high accuracy rates for both direct (90%) and indirect (89%) prompts&lt;/li&gt; 
 &lt;li&gt;We share a framework to help brands prioritize inaccuracies in need of fixing, based on how important a claim is, whether it's an easy fix, and if it's objective or subjective&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;Do You Know What AI Is Saying About Your Brand?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;A few months ago, we asked AI models, “what do people say about working with Seer Interactive?” Most mentioned high account manager turnover, which was a surprise to us.&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/ai-brand-accuracy-study" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Brand%20Accuracy%20Study-1.png" alt="Study:&amp;nbsp;AI Declined to Answer Prompts About Our Brand 31% of the Time (But Got Our Info Mostly Right When It Answered)" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h3&gt;TL;DR&lt;/h3&gt; 
&lt;ul&gt; 
 &lt;li&gt;We realized AI platforms were relaying an inaccurate narrative about Seer Interactive, so we ran a full analysis to find out what else they were saying — and whether the info was accurate&lt;/li&gt; 
 &lt;li&gt;Of the 6 AI platforms we tracked, they averaged a 95.9% accuracy rating when they answered the prompt&lt;/li&gt; 
 &lt;li&gt;The AI platforms were more likely to decline to answer a prompt (31% on average) than to provide an incorrect answer about Seer (3%)&lt;/li&gt; 
 &lt;li&gt;AI frequently declined to answer prompts about comparisons (80%), but had high accuracy rates for both direct (90%) and indirect (89%) prompts&lt;/li&gt; 
 &lt;li&gt;We share a framework to help brands prioritize inaccuracies in need of fixing, based on how important a claim is, whether it's an easy fix, and if it's objective or subjective&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;Do You Know What AI Is Saying About Your Brand?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;A few months ago, we asked AI models, “what do people say about working with Seer Interactive?” Most mentioned high account manager turnover, which was a surprise to us.&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fai-brand-accuracy-study&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Author: Nick Haigler</category>
      <category>AI Study</category>
      <category>Author: Bryan Gunawan</category>
      <pubDate>Thu, 27 Aug 2026 18:59:59 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/ai-brand-accuracy-study</guid>
      <dc:date>2026-08-27T18:59:59Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>How we built an agentic loop to improve every skill in our company, instantly</title>
      <link>https://www.seerinteractive.com/insights/how-we-improve-the-ai-behind-the-workk</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/how-we-improve-the-ai-behind-the-workk" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Did%20you%20even%20notice%20(3).png" alt="How we built an agentic loop to improve every skill in our company, instantly" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2 style="line-height: 1.2;"&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li style="line-height: 1.75;"&gt;&lt;span&gt;I reviewed our &lt;strong&gt;traffic-decline-analysis skill&lt;/strong&gt;'s and found four real problems: a traffic number from the wrong source, a real drop mislabeled "stable," no way to trace where a number came from, and no long-term trend view&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;I gave that feedback in a 9-minute recording and an agent picked up my feedback and fixed it the same day, and &lt;strong&gt;it shipped to everyone at Seer.&lt;/strong&gt;&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;I found another problem&amp;nbsp;the skill was picking which checks to run instead of running all of them, every time&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;An offhand idea from that same review became a new feature days later. The tool now runs off a simple spreadsheet export instead of needing live client data&lt;/span&gt;&lt;br&gt;&lt;span&gt;&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;p style="line-height: 1.2;"&gt;&lt;span&gt;Here’s how it all happened…&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/how-we-improve-the-ai-behind-the-workk" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Did%20you%20even%20notice%20(3).png" alt="How we built an agentic loop to improve every skill in our company, instantly" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h2 style="line-height: 1.2;"&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/h2&gt; 
&lt;ul&gt; 
 &lt;li style="line-height: 1.75;"&gt;&lt;span&gt;I reviewed our &lt;strong&gt;traffic-decline-analysis skill&lt;/strong&gt;'s and found four real problems: a traffic number from the wrong source, a real drop mislabeled "stable," no way to trace where a number came from, and no long-term trend view&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;I gave that feedback in a 9-minute recording and an agent picked up my feedback and fixed it the same day, and &lt;strong&gt;it shipped to everyone at Seer.&lt;/strong&gt;&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;I found another problem&amp;nbsp;the skill was picking which checks to run instead of running all of them, every time&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;An offhand idea from that same review became a new feature days later. The tool now runs off a simple spreadsheet export instead of needing live client data&lt;/span&gt;&lt;br&gt;&lt;span&gt;&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;p style="line-height: 1.2;"&gt;&lt;span&gt;Here’s how it all happened…&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fhow-we-improve-the-ai-behind-the-workk&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Inside Seer</category>
      <category>author-wil-reynolds</category>
      <category>AI</category>
      <category>author-jordan-strauss</category>
      <category>AI Lab</category>
      <pubDate>Thu, 20 Aug 2026 18:48:49 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/how-we-improve-the-ai-behind-the-workk</guid>
      <dc:date>2026-08-20T18:48:49Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>What Is AI Saying About Your Brand? Develop a Brand Canon and Find Out.</title>
      <link>https://www.seerinteractive.com/insights/brand-canon</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/brand-canon" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Brand%20Canon.png" alt="What Is AI Saying About Your Brand? Develop a Brand Canon and Find Out." class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h3&gt;&lt;strong&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/strong&gt;&lt;/h3&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;span&gt;Brands are measuring AI visibility without tracking if the perception of their brand is even accurate&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;We recommend building a brand canon, which tracks facts, differentiators, brand claims, what your brand wants to be known for, and what you don’t&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Once you’ve created your canon, you can use it to compare what you want AI to say about your brand versus what it actually says (and fix the output by tracing claims to their source)&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Learn how to build your own brand canon, how to measure the accuracy of claims made by AI, and how to prioritize fixing falsehoods and inaccuracies&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;What Happens if AI Shares Inaccurate Info About Your Brand?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;A potential customer asks AI what your company does, and the answer comes back about 70% right (per Seer's analysis of 28,123 responses about our brand across 6 AI models). But nearly 30% of the information is wrong or outdated, or the model declines to answer. &lt;br&gt;&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/brand-canon" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/Brand%20Canon.png" alt="What Is AI Saying About Your Brand? Develop a Brand Canon and Find Out." class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;h3&gt;&lt;strong&gt;&lt;span&gt;TL;DR&lt;/span&gt;&lt;/strong&gt;&lt;/h3&gt; 
&lt;ul&gt; 
 &lt;li&gt;&lt;span&gt;Brands are measuring AI visibility without tracking if the perception of their brand is even accurate&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;We recommend building a brand canon, which tracks facts, differentiators, brand claims, what your brand wants to be known for, and what you don’t&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Once you’ve created your canon, you can use it to compare what you want AI to say about your brand versus what it actually says (and fix the output by tracing claims to their source)&lt;/span&gt;&lt;/li&gt; 
 &lt;li&gt;&lt;span&gt;Learn how to build your own brand canon, how to measure the accuracy of claims made by AI, and how to prioritize fixing falsehoods and inaccuracies&lt;/span&gt;&lt;/li&gt; 
&lt;/ul&gt; 
&lt;h2&gt;&lt;span&gt;What Happens if AI Shares Inaccurate Info About Your Brand?&lt;/span&gt;&lt;/h2&gt; 
&lt;p&gt;&lt;span&gt;A potential customer asks AI what your company does, and the answer comes back about 70% right (per Seer's analysis of 28,123 responses about our brand across 6 AI models). But nearly 30% of the information is wrong or outdated, or the model declines to answer. &lt;br&gt;&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fbrand-canon&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>Analytics</category>
      <category>author-john-lovett</category>
      <category>AI</category>
      <pubDate>Mon, 17 Aug 2026 16:37:16 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/brand-canon</guid>
      <dc:date>2026-08-17T16:37:16Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
    <item>
      <title>27 GEO RFP questions you should be asking in 2026</title>
      <link>https://www.seerinteractive.com/insights/geo-rfp-guide</link>
      <description>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/geo-rfp-guide" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/27+%20GEO%20RFP%20questions%20you%20should%20be%20asking%20in%202026.png" alt="27 GEO RFP questions you should be asking in 2026" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;p&gt;&lt;span&gt;We looked back at hundreds of sales calls about GEO and looked for the questions people are asking, the pressures they are under, and decided to write up the questions people are asking and the questions they should be asking in an RFP situation.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Like my &lt;a href="https://www.seerinteractive.com/insights/seo-rfp-guide"&gt;SEO RFP guide&lt;/a&gt;, this guide isn’t limited to those of you looking for an RFP template for your upcoming GEO/AEO project: &lt;/span&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;br&gt;&lt;/span&gt;&lt;/p&gt;</description>
      <content:encoded>&lt;div class="hs-featured-image-wrapper"&gt; 
 &lt;a href="https://www.seerinteractive.com/insights/geo-rfp-guide" title="" class="hs-featured-image-link"&gt; &lt;img src="https://cdn.seerinteractive.com/hubfs/27+%20GEO%20RFP%20questions%20you%20should%20be%20asking%20in%202026.png" alt="27 GEO RFP questions you should be asking in 2026" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"&gt; &lt;/a&gt; 
&lt;/div&gt; 
&lt;p&gt;&lt;span&gt;We looked back at hundreds of sales calls about GEO and looked for the questions people are asking, the pressures they are under, and decided to write up the questions people are asking and the questions they should be asking in an RFP situation.&lt;/span&gt;&lt;/p&gt; 
&lt;p&gt;&lt;span&gt;Like my &lt;a href="https://www.seerinteractive.com/insights/seo-rfp-guide"&gt;SEO RFP guide&lt;/a&gt;, this guide isn’t limited to those of you looking for an RFP template for your upcoming GEO/AEO project: &lt;/span&gt;&lt;span style="white-space-collapse: preserve;"&gt;&lt;br&gt;&lt;/span&gt;&lt;/p&gt;  
&lt;img src="https://track.hubspot.com/__ptq.gif?a=7982212&amp;amp;k=14&amp;amp;r=https%3A%2F%2Fwww.seerinteractive.com%2Finsights%2Fgeo-rfp-guide&amp;amp;bu=https%253A%252F%252Fwww.seerinteractive.com%252Finsights&amp;amp;bvt=rss" alt="" width="1" height="1" style="min-height:1px!important;width:1px!important;border-width:0!important;margin-top:0!important;margin-bottom:0!important;margin-right:0!important;margin-left:0!important;padding-top:0!important;padding-bottom:0!important;padding-right:0!important;padding-left:0!important; "&gt;</content:encoded>
      <category>author-wil-reynolds</category>
      <category>GEO</category>
      <pubDate>Mon, 10 Aug 2026 18:37:44 GMT</pubDate>
      <guid>https://www.seerinteractive.com/insights/geo-rfp-guide</guid>
      <dc:date>2026-08-10T18:37:44Z</dc:date>
      <dc:creator>Admin</dc:creator>
    </item>
  </channel>
</rss>
