How do you check your brand's visibility in ChatGPT?
The obvious way — opening ChatGPT and asking about your category — is also the least reliable. Answers vary between sessions, your own chat history can color what you see, and a single question tells you nothing about the dozens of ways buyers actually phrase it. A real check needs the same set of buyer questions, asked cleanly, with every answer recorded and counted. That is the difference between an anecdote ('ChatGPT mentioned us once') and a measurement ('ChatGPT names us in 6 of 10 buying questions, always behind two competitors').
For each answer, MaxAEO records three layers. First, the brands: who is named, in what order, and whether you are recommended or merely mentioned in passing. Second, the framing: whether ChatGPT describes you positively, neutrally or with caveats. Third, the sources: for most product questions ChatGPT runs a web search before answering, and the pages it reads — comparison articles, community threads, category listicles — are the pages that shaped who it recommended. The sample above shows all three layers on real Slack data, including two questions where even Slack goes unmentioned.
The results point directly at the work. If ChatGPT cites a comparison page that omits you, that page is your target; if it leans on a community thread, that conversation is where your users' voices are missing; if it answers from memory with an outdated description, the fix is publishing a clearer, current account of what you do. Re-run the same questions after each content push and after major model updates — both move the numbers.