Picture a client who asks whether they show up in ChatGPT. Someone on the team types the client's name, sees it in the answer and sends a screenshot. A week later a competitor's screenshot shows the opposite. Both are real, and neither is a measurement. AI answers are generated fresh every time. The honest method is the same questions, worded the same way, asked on a schedule, with the sources written down.
Why one check proves nothing
Every answer is composed on the spot from whatever the search step returned, and the model's wording and choice of sources shift between runs. Personalisation from a logged-in account's history, the searcher's location, the model version and whether the search step fired at all all push the result around, so two people asking one question can get different citations. None of the companies documents how much it varies. Anyone who repeats a question will see it.
Three rules follow. Ask the same wording every time. Ask in a clean state: a fresh chat, logged out where you can be, and set to the market you sell in. Then repeat on a schedule and read the trend rather than any single result.
A prompt that contains your brand name ("tell me about Harbour Dental") will always find you, so it proves nothing. Use the questions a customer asks before they know your name.
Separate "an AI answer exists" from "it cites you"
These are two facts and they need two columns. For any query there are three outcomes. Sometimes no AI answer appears, the blue links carry the query, and there is nothing to lose. Sometimes an answer appears and cites other sites, which is the gap worth working on. And when an answer cites you, the cited pages are doing the work and deserve protecting.
Many transactional and local queries show no AI Overview at all. Treat that as a genuine zero, not a defeat.
When the citation goes elsewhere, write down which domains got it. For recommendation questions they are often roundups, directories, forums and review sites rather than your direct rivals. Those domains show you where to earn a mention.
Build a baseline you can repeat
A spreadsheet is enough. The hard part is keeping it identical from month to month.
- Write the questions a customer asks before they know your name. Mix the sub-questions from the free AI query fan-out simulator with the questions your sales conversations start with, and keep the list short enough that you'll really rerun it every month.
- Pick the assistants that matter for your market. Google's AI Overviews and AI Mode, ChatGPT, Gemini and Perplexity cover most businesses.
- For each question and each assistant, record whether an AI answer appeared, whether your brand was mentioned in the text, whether your site was cited with a link, and which domains were cited.
- Run your closest competitors through the same questions in the same session, so the comparison shares conditions.
- Repeat monthly with the same wording and settings. Add a column instead of editing old ones.
The payoff comes when "appeared, not cited" rows turn into "cited" after the work from the writing lesson.
What your analytics can and can't tell you
Google's Search Console Help says impressions and clicks from AI Overviews and AI Mode are included in the Performance report, but not broken out as a separate filter. Your Search Console data already contains that traffic, and you can't isolate it there.
In your analytics, a click on a citation arrives as a referral from the assistant's own domain. You'll see chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com in referral reports, so group them into one channel and watch them together. Read the count as a floor. Not every assistant visit carries a referrer, and a click from an AI Overview looks like any other Google organic click.
Mentions of your brand across the indexed web lead the rest, because assistants cite third-party pages when someone asks for a recommendation. Count mentions of your distinctive brand phrase and watch which way the number moves between checks. If your name is a common word, measure the distinctive part of it; otherwise most of what you count is noise. The Brand mentions and entity SEO lesson covers how to grow that number.
The ways this measurement lies
- A citation that shows up in one run and vanishes in the next is noise until it repeats.
- A mention in the text and a citation with a link are different outcomes, so report them separately.
- A query with no AI answer isn't a loss. A report that counts it as one will send you after the wrong keywords.
- Don't credit last week's change for a citation unless the same query, in the same wording, was uncited the month before.
Frequently asked questions
What is AI visibility?
It's how often AI assistants mention your brand in an answer or cite your site with a link. Those are different outcomes, so you track them in separate columns. Visibility only means something for the questions a customer asks before they know your name.
How can I check my AI visibility?
Don't rely on one search. Write a short, fixed list of customer questions, ask each in a clean chat on every assistant that matters for your market, and record whether an answer appeared, whether you were mentioned, whether you were cited and which domains were. Rerun it monthly with identical wording and read the trend.
What is AI visibility diagnosis?
Diagnosis means working out why you're missing from an answer. Start with the three outcomes: no AI answer, an answer that cites others, or an answer that cites you. The "appeared, not cited" rows are the ones to investigate, and the domains cited instead of you show where to earn a mention.
What is a good AI visibility score?
The lesson doesn't define a universal score, and no assistant publishes one. A useful benchmark is your own baseline and your closest competitors' results on the same questions in the same session. Judge yourself by whether your cited rows grow over several months.
What to take away
- A single screenshot of an AI answer is not a measurement. Answers vary by run, account, location and model, so only repeated checks with fixed wording count.
- Track whether an AI answer exists and whether it cites you as separate facts, because "no AI answer" is a neutral result.
- Search Console folds AI Overview and AI Mode traffic into the Performance report without a filter, and referral reports show assistant clicks only when a referrer is passed. Both are floors.
- Which third-party domains get cited instead of you is the most useful column in the sheet, because it tells you where to earn a mention.
Next
The domains an assistant cites instead of you are usually sites that link to and talk about the businesses they recommend, so the next chapter starts with Backlinks that matter, and the ones that don't.