Does ChatGPT name your brand?
The AI answer checker reports which buyer prompts name you, which name a competitor, and where you sit in the citation order across ChatGPT, Perplexity, Gemini and Google AI Overviews.
The answer engine is in build.
Today the checker returns the report shape filled with sample data, labelled as sample on every panel. No number on it describes your domain. The manual method below works right now, and the technical checks that decide whether an engine can cite you at all are live and free.
What Is an AI Answer Check?
An AI answer check measures whether an AI assistant names your brand when someone asks it a buying question in your category, and which domains it cites ahead of you. It is a different measurement from rank tracking. A rank tracker tells you your position in a list of links. An answer check tells you whether you appear in the one paragraph that replaces the list, and whether you appear first, third, or not at all.
Position matters more here than it does in search results. An assistant names two or three sources and stops. There is no page two to be found on, and a buyer who acts on the answer never sees the sources that were not named.
For the technical preconditions, whether the engines can reach and read your site at all, use the website audit or the AI visibility checker. Both run live and neither depends on the engine that is still in build.
How to Check It by Hand
Until the engine runs it for you, run it yourself. Three prompts, four engines, twenty minutes. This is the same measurement, done at a smaller scale.
- 01Write three prompts your buyer would typeNot keywords. Whole questions, in the words a customer uses. One discovery question, one trust question, and one head to head comparison naming two competitors. Include the category and the place if the purchase is local.
- 02Run each one in all four enginesChatGPT, Perplexity, Gemini and Google AI Overviews. The same prompt returns a different set of sources in each, which is why one engine is not a measurement.
- 03Record every cited domain, in orderPosition is the whole game. Note which domain is named first, second and third, and whether yours appears anywhere. Repeat the pass monthly and the trend tells you whether your fixes landed.
The three shapes that work: "best [category] for [need]", "where can I buy [product] with [condition]", and "[competitor A] or [competitor B], which is better for [use case]". Swap in the words your own buyer would type, or the answer describes somebody else's market.
What the Engine Will Report
A fixed buyer prompt set run across the four assistants, with the domains each answer cited and the order it cited them in. Your share of answer next to the competitors named ahead of you, per engine and per market, re-run on a schedule so the change is visible rather than claimed.
It arrives with the numbers or it does not arrive. There is no delivery date on this page, because we do not have one to give you.
In the meantime the work that moves the outcome is access and evidence, and both are covered in the GEO overview.
Frequently Asked Questions
Is the AI answer checker live?
Not yet, and the product says so on every panel. The engine that runs buyer prompts through ChatGPT, Perplexity, Gemini and Google AI Overviews is in build. What the checker returns today is the report shape filled with sample data, labelled as sample wherever it appears, so you can judge the format before anything is charged for. No number on it describes your domain.
What will the checker measure when the engine is live?
For a fixed set of buyer prompts in your category it will record, per engine, whether your domain is named, which domains are cited ahead of you, and in what order. Citation order is the measurement that matters, because an assistant names two or three sources and stops. Being cited fourth is close to not being cited at all.
How can I check my AI visibility by hand today?
Write three prompts a real buyer would type in your category, then paste each one into ChatGPT, Perplexity, Gemini and Google AI Overviews. Write down every domain each answer cites and the order it appears in. That manual pass is exactly what the engine will automate at scale. It costs you twenty minutes and it tells you where you stand.
What can I measure right now instead?
The preconditions, and they are the ones that break most often. The website audit resolves which AI crawlers can reach your site, whether your primary content is in the HTML rather than painted by JavaScript, whether llms.txt exists, and whether your structured data states which organisation the pages belong to. Those checks are live, they are free, and a failure in any of them means no engine can cite you regardless of how good your content is.
Why not just show a score anyway?
Because a score with no engine behind it is a number somebody made up. Plenty of tools sample a handful of prompts once and present it as coverage. We would rather report what is measurable, label the sample as a sample, and ship the number when the engine can produce it.
An engine that cannot read you will never cite you.
Crawler access, content in the HTML, llms.txt and structured data are measurable today and they are where most brands are losing. Run the audit, then work the list from the top.