2025 - ChatGPT named 21 competitors & us 0 times 2026 - Them 0, us 65%

In 2025 ChatGPT rewarded our competitors for ranking themselves in listicles.

We bet on humans instead, this is the data on what happened.

In February 2025 I recorded this video:

Everybody was asking the new hot question: how do I show up in LLMs? And the popular answer was the nice one:

• Put out thought leadership

• Do PR

• Build your brand

Do these things and the AI will find you, not true in Feb 2025.

I remember hearing that “make great content” shizz back in the Google days, AI optimization had the same moment.

14 years ago, I wrote how Google makes liars out of the good guys in content marketing.

I'd been doing "great marketing" speaking about GEO everywhere: Brighton SEO, MarketingProfs, MNsearch, Friends of Search. Madrid, Amsterdam, Munich, Sao Paulo.

We had just landed a study in Search Engine Land that got picked up by TONS of trade publications.

And when I asked ChatGPT for an agency strong at the intersection of PPC and SEO (literally the thing we've been doing for over a decade), I realized that in spite of living that life for 10 years. We had No visibility. For that and several other prompts that I thought we legitimately had a chance to.

For proof, here is me joining PPC and SEO data on Twitter using tableau 10 years ago!

The problem in Feb 2025: agencies ranking themselves in listicles

I asked ChatGPT three prompts that I knew should trigger my agency at least sometimes:

• "Give me a bulleted list of agencies strong at the intersection of PPC and SEO strategy. Just the company names."

• "I'm looking for an innovative SEO agency."

• "What are the best data driven SEO companies"

Across those answers, ChatGPT named 21 agencies. Seer wasn't one of them. EVER.

Then I clicked through to what ChatGPT was reading. Over and over, it was the same move: an agency publishes a "best SEO companies" page, lists itself at or near the top, and fills out the rest with other agencies.

Same names, same pages, over and over. It was shocking because many of these agencies I had never even heard of once, and I fancy myself as a guy who gets out and about to conferences and reads a ton.

I made the bet that self published listicles wouldn’t last as the optimal way to show solutions in B2B.

I bet on authentic intelligence to get my visibility in artificial intelligence, and that cost me 21 months of low visibility.

So I said I was going to stick with what I called authentic intelligence: win with real humans, on places like LinkedIn, where "real people are voting with real clicks."

We did integrated, human-loving, PR-loving content and suddenly the humans at trade pubs started including us and our insights.

We found out what humans loved, took that shit on the road, presented it everywhere, and then crushed with even more humans.

It was a team effort, we have always wanted to use our platform to help all team members put their thoughts in the world.

In the last 12 months, 10 Seer researchers had their work picked up by 19 publications.

Between August 2025 and August 2026 we tracked 958 mentions of Seer Interactive, and every one of those credits is an outlet citing a Seer study or POV.

Then we took it on the road. Here's what that did with actual humans: newsletter signups by channel, November 2024 through October 22, 2026.

Organic Social performs 14x better per session than SEO at driving signups of our newsletter.

What does organic social also bring?  Trust from people.

This is with SEO including our branded search too. Interesting to see how referral is 3.5x better than search, and speaking is 21x better.

This is that whole "Seen, Believed and Chosen" mantra we've been saying.  Being seen in AI or Google are my lowest channels at the rate of being chosen.  Even AI search is only marginally better than Organic.

Speaking & Social are 11% of our traffic and 54% of our newsletter signups.

Organic search is almost half our traffic and 13% of our signups.

And AI assistants? 2,359 sessions and 9 signups over two years. So no, I'm not ditching humans to chase robots.

Feb 2025 vs Nov 2026 - Rerunning the prompts, AI is finally reading human signals better

These are prompts I picked specifically to see if we'd show up. Today I ran the exact same 3 prompts again 10 times each on three current ChatGPT models (gpt-6.1-sol, gpt-6-astra and gpt-5.6-terra), with web search on.

For each of the 21 agencies, I only counted the prompt it was named on in 2025. That gives each agency 30 chances to show up again.

21 agencies named, 0 stuck in the newest ChatGPT models, but Seer went from 0 to 58 / 90 runs.

Of the 21 agencies ChatGPT named in February 2025:

• 0 show up reliably (2 in 3 answers or more)
• 4 show up sometimes (Directive, Ignite Visibility, WebFX, SmartSites)
• 8 are rare (under 1 in 3 answers)
• 9 are gone completely: 0 mentions

Add up the bottom 17 and it gets stark. They had 540 chances to be named. They were named 27 times.

Now, remember these prompts were intentionally biased back in Feb 2025 to show Seer, they were ones that were strongly associated with our brand.

That's 5%. About 1 in 20.

Nine agencies that ChatGPT confidently recommended in 2025 didn't get named a single time across 270 tries.

In 2025 I said, "This can't be how to show up in AI forever." Here we are in November 2026, yeah it wasn't.

How we did across each of the three prompts...

• Agency strong at the integration of PPC + SEO strategy: 25 of 30 
• Data driven SEO companies: 28 of 30
• Innovative SEO agency: 5 of 30. More on that one below.

Listicles still work, just a LOT less.

Here's the honest part. I wanted the story to be "the listicle spam died and the good guys won." That's not quite what the data says.

When ChatGPT names Seer, it's still reading "best agencies" lists. In the 58 answers that name us, it cited Seer's own site in 21, and those were our homepage and our /work pages. We never built a "best agencies" page ranking ourselves.

So I went and read the four other pages it cited most in those answers:

Every one of them is the exact format I called out in 2025: an agency writes the list and puts itself first. That play didn't die. All four of these pages were updated in the last few months.

While we do the hard work of winning humans, publishing real stuff, those agencies are still playing the "loophole" game.  We ain't building for that future.

So the list game is still in the game, but as Lily Ray stated here, 70% of the time a listicle was cited, it recommended your competitors and NOT you. So the listicle worked AGAINST your visibility and positively for your competitors.

Where we still lose

I'm not going to pretend this is a clean sweep.

On "I'm looking for an innovative SEO agency," Seer shows up in 5 of 30 answers. iPullRank shows up in 23 of 30. Mike and that team earned it, and across all 90 answers they're in 50. 

Caveats, because you should ask

• The 2025 baseline is one run. That's me in ChatGPT on GPT-4o, recorded on video. Some of those 21 agencies may have been flickering in and out back then too. We just didn't measure it that way.
• The setup isn't identical. 2025 was my logged-in ChatGPT. 2026 is the same prompts through an API with no memory and no personalization, on newer models.
• I can't prove it line by line. I can't draw a straight line from a specific talk or LinkedIn post to ChatGPT naming us. But I know what we did, and I just showed you what it did with humans.
  
  

Who Seer is aiming to be...the agency that builds strategies that win in the short and long term.

A reckoning is coming. I see it in sales calls. In 2025, the fast play was to write a page ranking yourself and wait for the AI to read it.

A lot of the agencies decided to become loophole-ists and not strategists - they got their moment for their clients, short term wins worked, but the long term hangover is coming... now their clients are addicted to "fix this in 4 weeks" vs I am willing to do long term work to align my brand and what AI thinks about me.

21 months later, most of those names are gone from the same answers.

In the words of Kool Moe Dee... How Ya Like Me now?

While we were getting crushed, in spite of trying so hard to write content and share insights that real people, humans would love, we made the bet that this work would pay off. It was already paying off in signups for our newsletter and pipeline. Now it is starting to in AI too.

Now we just need to double down on sharing what we learn, which has been our ethos since day 1, we're just finally being rewarded for it.

We took the slow road. We did integrated, human-loving work, took it on the road, and let other people do the talking. It worked with humans first. Then the AI caught up.

We can win for both AI and people. That's the GEO offering we want to build.