08/27/2026
AI loves finding negative things to say about your brand.
When AI considers/recommends a brand, the answer usually has a "but." Great option, but availability varies. Solid product, but pricing complaints.
We specifically track those "buts" across AI platforms - ChatGPT, Perplexity, Gemini, Copilot, and Google's AI surfaces. This week I pulled where the "but" influence comes from by pulling query fanouts and AI answers across ChatGPT, Gemini, and Perplexity for around 7,500 "but answers" over two weeks.
Every industry has its own judgment layer, and that's who seems to influence the narrative the most right now.
For a healthcare provider it was Yelp and BBB. For a consumer app, the App Store and Google Play. For a hardware company, a random complaints website and a hobbyist forum. For our own agency, Clutch and DesignRush.
BBB and Trustpilot show up almost everywhere because those also show up in ChatGPT site: query fanouts more since August 8th. In our systems, on ChatGPT, Reddit now appears on only ~6% of "but answers". But on Perplexity, Gemini, Copilot, and Google's AI surfaces, Reddit is still on roughly half of them. Reddit didn't stop influencing the "but". It stopped influencing them in one place. This is why it's important for AEO/GEO to have understanding of what LLMs and AI surfaces your target audiences are using and what they're using them for.
Two specifically interesting pieces of info:
A separate healthcare brand's warnings were influenced most by Indeed and SimplyHired in about a third of its answers. Employee reviews are influencing patient-facing recommendations because other sources didn't really have negative things to say!
And one brand's "but answers" cited a direct competitor's website half the time. Not because the competitor website is an authority, instead the competitor website is filling a gap beause no authoritative source has the "but" information about these brands/industry.
