AI search visibility: how to check if ChatGPT mentions your brand

To check your brand's visibility in AI search, ask ChatGPT the questions your buyers ask, in their words, several times in a fresh chat, and note whether you are named, where you appear and which sources are cited. One answer proves little, because answers vary; a set of real buyer questions, asked repeatedly and tracked over weeks, shows your true position.

This guide walks through the manual check, what to record, how to read the results, and when it is worth automating.

Why a single question is not enough

AI assistants do not return a fixed list like a search results page. Ask the same question twice and you may get different brands, in a different order. Phrasing matters too: "best invoicing app" and "invoicing software for freelancers who bill in two currencies" can produce completely different shortlists.

So the honest answer to "does ChatGPT recommend us?" is a percentage, not a yes or no. The goal of the check is to estimate that percentage for the questions that lead to sales.

Step 1: Write down the questions your buyers ask

Start with 10 to 25 questions. Good sources:

• Sales calls and demos. What did prospects ask before they found you?
• Support tickets and onboarding notes. What problem were they trying to solve?
• Search Console. Long, question-style queries that already bring you visitors.
• Competitor comparisons. "[Competitor] alternatives" and "[You] vs [Competitor]".

Mix three kinds of question:

Leave out questions that contain your brand name. "Is Acme good?" tells you how ChatGPT describes you, not whether it recommends you.

Step 2: Ask in a clean session

Your chat history and saved memories can shape answers. To see what a new buyer sees:

1. Open a temporary chat, or log out.
2. Paste each question exactly as a buyer would type it.
3. Ask each question at least three times, in separate chats.
4. Repeat in Gemini, and search the question on Google to see whether an AI Overview appears.

If ChatGPT searches the web for the answer, it shows the sources it used. Those are the most useful part of the check.

Step 3: Record four things for every answer

A simple spreadsheet is enough. For each answer, note:

• Named: is your brand mentioned at all?
• Position: first, second, fifth, or only in passing?
• Competitors named: which brands appear instead of you?
• Sources cited: which pages did the engine link to?

After 25 questions asked three times, you have 75 answers. Your visibility share is the percentage of answers that name you. Do the same for each competitor, and you have a clear picture of who owns the shortlist.

Step 4: Read the gaps

The results usually fall into one of three patterns.

You are not named anywhere

The engine does not know enough about you, or cannot find pages that describe you for these questions. Start with the basics: make sure AI crawlers can read your site, that your home page states plainly what you do and for whom, and that you are listed on the directories and review sites in your category. Our free AI readiness checker shows which crawlers your site blocks and what is missing.

You are named, but rarely and low down

The engine knows you but trusts other sources more. Look at the cited pages. If the same comparison article, review site or Reddit thread keeps appearing and it lists your competitors but not you, that page is the gap. Getting onto it, or publishing a better answer yourself, is the fix.

You are named for some questions and not others

This is the most common result, and the most useful. It shows exactly which buyer questions you are losing. Write a page that answers each lost question directly, with the specifics a buyer needs, and link to it from your main pages.

Step 5: Re-check every week

Answers change as engines update and as new pages are published. A one-off check is a snapshot; the trend is what tells you whether your work is paying off. Re-run the same questions on the same day each week, and compare your visibility share and the cited sources with the week before.

This is also where the manual method starts to hurt. 25 questions, three times each, across three engines is 225 answers a week to read and log.

Checking automatically

You can check one question for free with the CiteGaps free check: it asks ChatGPT a question your buyers ask and shows whether you are named, with no account needed.

CiteGaps then does the full routine for you. It asks ChatGPT, Gemini and Google's AI Overviews your buyer questions every week (Starter and Agency add Perplexity and Google AI Mode), works out your visibility share and citation share against competitors, lists the pages each engine cited, and drafts the content to close the gaps it finds. If you want to understand the metrics first, read what AI answer-engine visibility is and how to measure it.

Questions

Is there a free AI visibility checker?

Yes. You can ask ChatGPT yourself in a temporary chat, or use the CiteGaps free check, which runs one buyer question and shows whether you are named, with no sign-up.

How many questions should I check?

Start with 10 to 25 real buyer questions and ask each one at least three times. Fewer than that and a single unusual answer can mislead you.

Why does ChatGPT mention my competitor and not me?

Usually because the pages it trusts for that question mention your competitor and not you: directories, review sites, comparison articles and community threads. Check the sources it cites to find which pages to fix.

How long until AI answers change?

Answers that rely on web search can change within weeks of a new page being published and found. Answers based on what the model learned in training change more slowly, as new versions of the model are released.