September 2026 Insights and Action Items
1. Review and trust sites are the number two citation source in AI answers, and they take over right when people decide.
What's happening: Seer's analyzed more than 800,000 responses across ChatGPT, Gemini, AI Mode, and Perplexity and found review sites make up just 3% of citations at the awareness stage, where Reddit and editorial sites dominate discovery, then climb to 23.5% as users move down the funnel toward a decision.
2. Having no reviews is its own signal, and it costs you in head-to-head comparisons.
What's happening: Models will state outright that a brand has "a distinct lack of extensive verified third-party reviews," and in comparison prompts the brand with more reviews had its evidence pulled into responses nearly twice as often as the brand with fewer, even when both had substantial review inventories.
3. One prompt run shows you less than half the field, so three runs is the floor.
What's happening: Seer's prompt frequency test across just under 100 prompts found a single run surfaces five or six brands where ten runs surface twelve or thirteen, with three runs getting you a little more than 70% of the brands ChatGPT will ever name — and legal, beauty, health and fitness, and energy categories need more than that.
4. If you're the first brand named, that position is stickier than everything below it.
What's happening: The usual complaint about AI search is that the same prompt reshuffles the same brands every time, but 75% of runs repeated the same lead recommendation — a spot that's typically inherited from existing market leadership rather than earned through a GEO program, so treat it as a benchmark to defend.
Action Items
1. Pull the review platforms getting cited for your brand's evaluation prompts, then claim those profiles and monitor themes and sentiment. How old is the review driving your brand's perception?
2. Run your 20 to 30 most important prompts ten times each and classify results as locked in, mixed, or volatile. Which of your current reports are built on a single run?
AMA with Alisa and Nick
Alisa and Nick's Latest Thoughts
Google Is Not Going to Stop Ranking Your Claude-Assisted Content
TL;DR 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 It’s a…
Study: AI Declined to Answer Prompts About Our Brand 31% of the Time (But Got Our Info Mostly Right When It Answered)
TL;DR 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 — …
Our Early View of ChatGPT Ads: 5 Initial Observations - Updated July 28, 2026
When we first published this on July 15, we were working from three days of ChatGPT ads data. Revisiting the data this week, our main takeaways haven't changed,…
The Proof: ChatGPT 5.5's fanout patterns reveal the importance of brand
If you've been looking at the whole "do I have to do listicles" thing and throwing up in your mouth a little, this post is for you.