You open ChatGPT. Type in your service and your city. Hit enter.
Then you scan the answer looking for your name.
More and more business owners are doing exactly this. And what they find varies wildly. Some appear immediately, with accurate information, recommended by name. Others find a competitor they’ve barely heard of. Some aren’t there at all.
This article helps you find out where you stand. You need four browser tabs and about ten minutes.
Why every AI tool gives you a different answer
Before you run the test, there’s one thing worth understanding: you’ll probably get different answers in different places. And if you run the same search twice on the same platform, the answer can change there too.
That’s not a bug. It’s how AI search works.
ChatGPT, Perplexity, Google AI Overviews and Microsoft Copilot use different training data, pull from different sources and weigh information differently. And even the same tool answers differently from one run to the next. A study from SparkToro and Gumshoe.ai in January 2026 showed it: volunteers put the same twelve brand questions to ChatGPT, Claude and Google’s AI, over and over.
What came out says a lot about how you should test. Almost every answer was unique: different brands, in a different order, in different numbers. The chance of getting the same list twice was under one percent, and the same list in the same order hardly ever turned up. A single test is a snapshot, not a verdict.
So a single result isn’t a verdict. It’s a signal. And that’s exactly why it’s worth checking all four.
If you want to understand what determines whether you get mentioned at all, and how AI visibility differs from traditional SEO, we cover that in our guide to GEO and AI visibility.
The test: four tools, four questions, ten minutes
Run the same four questions on each platform. Use an incognito window if you can, and make a note of what you find.
The four questions:
Which [your trade or service] in [your city] would you recommend?Who are the best [your industry] in [your city]?I need [describe the problem] in [your city], who do you recommend?What do you know about [your business name]?
Replace the brackets with your own details. A plumber in Bristol would ask: “Which plumber in Bristol would you recommend?” A bookkeeper in Edinburgh would ask: “Who are the best bookkeepers in Edinburgh?”
Here’s what to look for on each platform.
ChatGPT
ChatGPT is the most widely used AI tool and the first place most business owners check. Ofcom’s Online Nation report, published on 10 December 2025, found that 54 per cent of UK adults now use AI tools such as ChatGPT, Copilot or Gemini, and that ChatGPT alone drew 252 million UK visits in August 2025.
One thing is worth checking before you start. ChatGPT searches the web automatically when a question looks like it needs current information, but not every time — and without a search it answers from training data that can be considerably out of date. The interface changes often, so don’t go hunting for a particular button. Look at the answer instead: if there are no source links, it didn’t search. Rephrase the question or trigger the search manually.
Run your four questions and notice: does your name come up? Are you one name in a list, or does ChatGPT single you out? Is the information accurate — address, services, opening hours?
Then run the first commercial question again a few minutes later. Do you get a different answer? It’s more common than you’d expect. If competitors appear one time but not the other, it means you’re in the borderline territory of what ChatGPT knows about your category.
Perplexity
Unlike ChatGPT, Perplexity is a search engine built on AI. It always cites its sources, which makes it the most transparent tool for this kind of diagnosis.
Run the questions, but don’t just read the answer. Click through to the sources. Is your website among them? If a competitor appears and you don’t, check where Perplexity is pulling from. Is it the competitor’s own site, or a third-party list like “best plumbers in Bristol” on a comparison site? That distinction matters a lot for what you can do about it.
Perplexity frequently surfaces articles and lists from third-party directories. If you’re not listed on platforms like Trustpilot, Checkatrade, Yell or a sector-specific directory, that may be a direct reason why you’re being left out.
Google AI Overviews
Google AI Overviews is the AI-generated answer block that appears at the top of Google search results, above all the organic links. It’s tied directly to Google’s own index, which means your Google Business Profile and your website content have a direct effect on what appears.
Open google.com in an incognito window. This matters: logged-in results are personalised, which distorts the picture. Search for best [trade] in [city] and [trade] in [city] reviews.
Bear in mind that AI Overviews doesn’t always appear. For some searches, Google decides an AI summary isn’t warranted. What Google’s own documentation on AI features in Search is clear about is the requirement for being cited at all: a page has to be indexed and eligible to appear in Search with a snippet. There are no separate requirements beyond that. If an overview does appear and you’re not in it, note which sites are cited. Those are the sources Google currently considers most authoritative for your category.
Microsoft Copilot
Copilot is Microsoft’s AI search tool, integrated with Bing. It draws a fraction of ChatGPT’s traffic, and it’s almost always the last place businesses think to look. But it’s worth checking for a specific reason: Copilot indexes via Bing, not Google. If you’ve focused all your visibility efforts on Google and haven’t thought about Bing, Copilot is often where that gap shows up first.
Run the same questions. Copilot tends to produce more structured, list-based answers than ChatGPT. Do you appear? And are you being cited from your own website, or from a directory or mention on another site?
What the results actually tell you, and what they don’t
You’ve run the test. Here’s an honest read of what it means.
If you appear on all four, with accurate information and as an active recommendation rather than a name in passing, that’s a good sign. It doesn’t mean your visibility is strong, but it means you don’t have obvious gaps.
If you appear on some platforms but not others, that’s common for smaller local businesses. Each platform draws from different sources. The gaps tell you where the weaknesses are.
If you don’t appear at all, or a competitor consistently appears where you’d expect yourself, that’s a clear signal that something is missing. Usually it comes down to some combination of thin presence in external directories, missing structured data on your website, or content that isn’t written in a way AI tools can extract a clear answer from. Structured data, or schema markup, is one of the most practical ways to help AI tools understand what you offer and where you’re based. We explain what that means in our guide to schema markup.
If you show up but with the wrong details, the error almost always comes from one of three places: an outdated Google Business Profile, inconsistent details in directories such as Hitta.se, Reco and Allabolag, or old schema markup on the website. Start with the profile and check that name, address, phone number and opening hours match exactly. Then compare with the directories.
Note how you’re mentioned, too. Being mentioned means your name appears in the answer. Being cited means the tool uses your website as a source. Being recommended means you’re put forward as the answer. The difference is large: BrightEdge analysis from August 2025 found that ChatGPT mentions brands 3.2 times more often than it cites them.
But whatever you find, there are three things this test can’t tell you.
The first is whether your result is representative. You ran a handful of queries. A systematic audit runs many more prompts per platform, each one several times, phrased the way your actual customers ask questions. Your test is a sample.
The second is whether the results hold over time. Profound, a company that specialises in AI visibility measurement, compared around 80,000 prompts per platform in July 2025 against the same prompts a month earlier. Between 40 and 60 per cent of the cited domains had been replaced in that single month: 59.3 per cent for Google AI Overviews, 54.1 for ChatGPT, 53.4 for Copilot and 40.5 for Perplexity. Over six months, 70 to 90 per cent of the domains were different from where they started. What you see today could look quite different in six weeks, and there’s no way to know that from a single snapshot.
The third is how you compare to your competitors. “I appear sometimes” means nothing if your nearest competitor appears consistently in every answer. Visibility is relative. Without a comparison, you don’t have a real measure.
The test gives you a starting signal. If things look blank across all four platforms, you know enough to act. If it’s mixed — which it is for most businesses — you have a feeling but not a picture.
What a proper audit adds
The difference between what you just did and a systematic AI visibility analysis isn’t effort. It’s method.
A proper audit runs a large set of prompts, phrased the way your customers actually search, across all four platforms, in separate sessions without personalisation. It distinguishes between being mentioned, cited and recommended — three things that look similar in an AI answer but require completely different responses. It compares your results against two or three competitors so you understand where you actually stand. And it maps what’s causing the gaps: missing schema markup, weak third-party presence, inconsistent business information, an outdated Google Business Profile, or content that isn’t structured for AI extraction.
That’s what our AI visibility audit does, and we explain how the method works in a separate article.
Save what you noted in the test, with the date. If you order the audit, that’s your starting point: we run the same four questions you did, plus the questions your customers actually ask, and show how the picture changes with a larger sample.