To find out whether ChatGPT and Google AI mention your business, ask each engine the buying questions your customers ask and record whether you are named. There is no ranking dashboard inside these tools, so asking them directly is the only reliable check.
The work of earning those mentions has a name, generative engine optimisation, and it is worth reading if you want the theory. This guide is the practical half: how to find out where you stand today, and what moves the needle.
How do I check whether ChatGPT mentions my business?
There is no dashboard inside ChatGPT that tells you your “AI ranking”. The only reliable way to know is to ask the engines the questions your customers ask, and read who gets named. Do it across the engines that matter now:
- ChatGPT
- Google AI Overviews
- Claude
- Perplexity
Ask each one your ten most important buying questions. Phrase them the way a real customer would (“who is the best plumber in Newcastle for emergency work?”), not as keywords. Record whether you are mentioned, and which competitors are. That list is your baseline.
One thing to watch: answers vary between sessions and can be personalised. Ask each question a couple of times, in a fresh session, so you are reading the pattern rather than a single roll of the dice.
The engines do not behave the same way
This is the part most guides skip, and it changes how you read your own results. We ran 10,872 probes across five engines from our own production system between 31 May 2026 and 25 August 2026, and they split into two clear groups.
| Engine | Probes | What it did |
|---|---|---|
| Google AI Overviews | 7,494 | Produced an AI Overview on 3,400, or 45.4% |
| ChatGPT | 913 | Answered every probe |
| Perplexity | 911 | Answered every probe |
| Claude | 792, since 20 June 2026 | Answered every probe |
| Gemini | 762 | Usable citation data on only 94; last probe 3 August 2026 |
The plain reading, as at August 2026: Google is the only one of the five that often declines to generate an AI answer at all. The others always answer something.
That matters for two reasons. First, if you check one Google query, see no AI Overview and conclude you have no AI visibility problem, you may simply have hit one of the questions Google chose not to answer. Second, a single blended “visibility score” across all engines hides this completely. You want the engines reported separately, because they are not measuring the same thing.
Gemini is the honest caveat in that table. Our usable citation data from it dried up and our last probe was on 3 August 2026, so it is currently off in our own roster rather than sitting stale in a chart.
What actually improves your odds
AI answers tend to name businesses that are:
- Clearly the answer. Pages that state what you do, where, and for whom, in plain language near the top.
- Quotable. Specific, checkable facts an engine can lift and attribute: a price, a date, a measured result. Vague claims give it nothing to work with.
- Consistent across the web. Your name, category and location matching across your site, your Google Business Profile and directory citations.
- Well regarded. Reviews and mentions the engines can find.
- Reachable. Not blocked by a firewall or robots rule that stops the crawler reading you. This is the one that catches people out: a Cloudflare rule added to keep AI training crawlers out is very easy to write too broadly, so it locks out Googlebot as well. Our free Cloudflare crawler check fetches your site as Googlebot and as a normal browser and tells you whether the two get different answers.
The second point is the one we underestimated on our own site. When we scored our own nine articles with our own content tooling in August 2026, they carried 10 checkable data points between them across roughly nine thousand words, and seven of the nine had none at all. Every one scored below our 55-point E-E-A-T threshold, in a range of 22 to 47. Over the same period our non-branded probes were cited zero times. We wrote up the whole result and what we changed.
Most of that list is mechanical, which means you can measure it rather than guess. The free agent readiness check reads your homepage the way an answer engine does and scores it on crawlability, structured data, schema validity, entity recognition, freshness and topical relevance. It will not tell you whether ChatGPT names you, but it will tell you whether ChatGPT could. The academy guide on structuring a page for AI answers covers the writing half.
Notice how much overlap there is with good local SEO and technical SEO. The work that wins the map pack is largely the same work that wins the AI mention: a Business Profile in the right category, consistent details across directories, and a page that plainly answers the question being asked.
Branded and non-branded are two different questions
Before you read your baseline, split it in two. “Does the engine know who we are?” and “does the engine recommend us?” are separate questions with very different answers, and mixing them produces a number that means nothing.
Our own split, over the 60 days to 25 August 2026, shows how wide the gap can be:
| Bucket | Probes | Times cited |
|---|---|---|
| Branded, questions containing our name | 68 | 14 |
| Non-branded, questions a buyer would type | 876 | 0 |
If you only ever probe your own brand name, you will get a comfortable number that tells you nothing about whether new customers can find you. The non-branded set is the one that matters commercially, and it is usually the one that hurts.
A simple worked example
Say you run a physiotherapy clinic and ask, “best physio in Geelong for sports injuries?” Perplexity names three clinics and you are not among them. You open the three it did name and notice each has a dedicated page for sports injuries, a Business Profile in the right category, and a stack of recent reviews.
Your fix list writes itself: a clear sports-injury page that answers the question, a Business Profile category and description that match, and a habit of asking happy patients for reviews. Re-ask the question a month later and you can see whether the mentions have started to shift your way.
Turn it into a habit, not a one-off
A single check is a snapshot. The value is in re-running it on a schedule so you can see mentions appear after you fix a page or tidy your Business Profile, and catch it when a competitor overtakes you.
Keep the questions fixed while you change the pages. If you change both at once, the next number is unreadable, which is the same discipline that makes SEO ROI measurable rather than assertable. Measured over time, AI visibility becomes a metric you manage, not a mystery.
Argrow runs these checks for you across ChatGPT, Google AI Overviews, Claude and Perplexity, tracks your AI visibility beside your organic and local rankings, and measures the lift when a change earns you a citation. See how AI visibility tracking works.
Frequently asked questions
How do I check if ChatGPT mentions my business?
Ask it the buying questions your customers would ask, like 'who is the best [your service] in [your area]?', and read whether you are named and who else is. Do it across ChatGPT, Google AI Overviews, Claude and Perplexity, and repeat it on a schedule so you can track changes rather than a single roll of the dice.
Is there a tool to track AI search visibility?
Yes. Argrow asks each engine your customers' questions on a schedule and records whether you are cited, tracked next to your organic and local rankings, with the lift measured when a change earns a mention. We ran 10,872 of those probes across five engines between 31 May and 25 August 2026, including on our own site.
Why is my competitor mentioned by AI and I'm not?
Usually because their page answers the question more directly, their business details are more consistent across the web, they have reviews the engine can find, or your site is harder for the crawler to read. All of those are fixable.
How often should I check AI visibility?
A single check is only a snapshot. Monthly is a sensible rhythm for most businesses, so you can see mentions appear after you make changes and catch it early when a competitor overtakes you. Keep the questions fixed between runs, or you cannot tell whether the number moved or the test did.
Does an AI Overview appear on every Google search?
No, and this catches people out. Across 7,494 Google AI Overviews probes we ran between 31 May and 25 August 2026, an AI Overview appeared on 3,400 of them, or 45.4%. ChatGPT, Perplexity and Claude answered every probe we sent them in the same window.