How to check if ChatGPT recommends your business (a 10-minute test)

Do not ask "what do you know about [my company]?" That tests fame, not sales. Ask what a customer asks when they do not know you yet. Use this pattern:

Category question: "What is the best [category] for [customer type]?" For example: "What is the best accounting software for a small construction company?"

Problem question: "I need to [solve problem] without [common constraint]. What should I use?"

Comparison question: "[Your brand] vs [competitor]: which is better for [use case]?"

Purchase question: "Which [category] should I buy if my budget is [realistic amount]?"

Validation question: "Is [your brand] a credible provider of [category]?"

Write them the way your customers talk, not the way your marketing does. If your customers say "app to track my crew's hours," do not ask about "workforce management solutions."

Open a new conversation for every question, with chat history and memory disabled or in a private window, so earlier context does not color the answers. Run your five questions in at least two systems, ideally three: ChatGPT, Gemini and Perplexity make a good spread, because they balance model knowledge and live web search differently.

One practical note: if the assistant offers to search the web, let it. You want to see both behaviors, what the model remembers about your market and what it finds when it looks.

For every answer, note:

1. 01

   Mentioned: does your brand appear at all?

2. 02

   Recommended: does it appear as a suggestion, or only in passing?

3. 03

   Accurate: is the description of what you do correct, or outdated and wrong?

4. 04

   Instead of you: which competitors appear, and does the answer cite any sources?

A simple spreadsheet with five questions and three systems gives you fifteen data points. That is enough to see a pattern.

You appear and are recommended accurately. Good position. The work now is keeping it and expanding to more questions.

You appear, but described wrongly. This is the misdescription pattern, and it quietly kills conversions: the assistant is selling a version of your company that does not exist. The fix is usually a mix of clearer pages and corrected third-party sources.

Competitors appear, you do not. The most common result. It usually means the evidence AI systems lean on, reviews, directories, comparison pages, independent descriptions, supports them and not you.

Nobody appears, answers are generic. Your category itself is weakly covered. That is an opportunity: the first business to build clear, citable evidence tends to become the default answer.

See the full picture in minutes: Run your AI Visibility Audit, $19 →

Be honest with yourself about the limits of this exercise, because they are exactly where wrong decisions get made:

One run is not a measurement. AI answers vary between runs. A brand can appear in one answer out of three and you happened to catch the one. Reliable measurement repeats questions and reports rates, "appeared in two of three runs," not verdicts.

You cannot see why. The manual test shows the outcome, not the cause. Whether the gap sits in your website's machine readability, your content coverage or your external citations requires checking each layer separately.

You cannot see the sources. When a system recommends your competitor, the reasons live in the citations. Without extracting and comparing them, "create more content" is a guess.

This is the exact gap the EntityRise AI Visibility Audit closes: it runs 15-25 buyer questions with repeated runs across ChatGPT, Claude, Gemini, Grok and Perplexity, records citations and competitors, scores four separate foundations and returns a prioritized fix plan. It costs $19 and takes minutes, so the sensible order is: run the manual test today, and if you see any of the three problem patterns above, measure properly before you change anything. The methodology is public.