How to check whether AI recommends you as an expert

Typing your name into ChatGPT can tell you whether it can describe you. A potential client who has never heard of you will ask a different question.

They might ask for an independent consultant who can fix an expensive Google Ads account. Or a fractional finance director with experience in subscription businesses. They have a problem and want someone who can help.

Start your check with that problem. Leave your name out of the question, save the answer, and look closely at who appears. You can do this manually in the AI tools you care about. The useful result is a record you can inspect and repeat.

Write the question before opening the chat

For an illustrative example, imagine Alex Morgan, a fictional consultant who helps US software companies with Google Ads. These two questions test different things:

What does Alex Morgan do?

Which independent consultants help US software companies improve their Google Ads campaigns?

The first supplies the name. The second asks the engine to find someone. Keep those results separate, even if both answers mention Alex.

Choose a small starting set that covers the expertise you want clients to hire you for. Include a location or an industry when it would affect the client's choice. Avoid packing your entire biography into the question. An obscure combination of details can make you the obvious match while telling you very little about a normal search for help.

If you have enquiries or notes from client calls, start with the language in those. Remove names and confidential information before entering it into an external service. Otherwise, write down the situation a client would need help with and treat the wording as a hypothesis to test.

Keep the test conditions in your notes

Use a fresh conversation for each manual check. If you first spend ten messages explaining your background, a later recommendation has that context available. It no longer answers the same question as a chat with someone who has never introduced you.

A fresh conversation still does not create a universal result. OpenAI's Search guide suggests specifying a location when it matters. Its Memory documentation explains that relevant remembered details can be used to refine a search query. Record whether you used search and any relevant account context. Other tools have their own settings. ChatGPT Search, Memory in ChatGPT

Save the exact question rather than a paraphrase. Also record the engine, date and any location you specified. When you check again, these notes help you see whether the answer changed or you asked a different question.

Gistful's scans are controlled samples across ChatGPT, Gemini, Claude and Perplexity. They should not be read as a replay of every client's personal chat. The methodology documentation explains that boundary.

Read beyond the name

A matching name is the beginning of the check. Open the supporting links, where the answer provides them. Does the website, role or location establish that this is you? Common names make this especially easy to get wrong.

Then read the sentence around the mention. The answer may recommend contacting you, quote an article you wrote, or discuss someone with a similar name. Those are different findings.

The guide to citations and personal recommendations covers these distinctions in more detail. Use a simple record like this:

This is a manual reading guide, not a replacement for a product's scoring definitions. Gistful's live coverage metric counts confirmed identity-supported mentions in returned discovery answers. Its results documentation explains the calculation.

A failed request cannot establish that you were absent. Keep it out of the denominator if you calculate your own appearance rate.

Follow one piece of evidence

Suppose the answer recommends another consultant and cites their page about software-company ad accounts. Open the page and examine what it says. Perhaps it describes the exact service in the question, explains who it serves and includes a detailed example of the work.

That gives you something concrete to review on your own site. It does not prove why the engine chose that person. A citation shows a source attached to the answer; it is not a complete explanation of the engine's selection process.

Compare like with like. If you do that work but your site only says “growth expert,” a more specific service description would help a reader understand your offer. If the cited person has experience you do not have, copying their positioning would misrepresent your work.

Pick an improvement you can support: clarify an existing service, explain a process you use, or publish a permitted example of completed work. Resist making a list of twenty new articles before you have read the evidence from the first question.

For a closer source review, see why AI recommends other experts instead of you.

Save the baseline before changing anything

Keep the original answers and links, even when the result is disappointing. A screenshot alone may cut off the question or sources, so save the text and the test conditions too.

After updating your public material, run the same questions again under comparable conditions. An unchanged answer on an immediate repeat does not tell you whether your new page has been discovered or considered. Nor can a later improvement, by itself, prove that your edit caused it.

Google notes that discovery and indexing are not guaranteed. Its guidance for AI search builds on ordinary search foundations, including accessible pages and useful original content. Google's AI search guidance

For your first check, choose one question you would want a prospective client to ask. Save the complete answer and verify every reference to your name. If you prefer to keep that evidence in Gistful, start with the question setup guide.