Key takeaways
- One check isn't enough: SparkToro found ChatGPT and Google's AI gave the same list of brands less than once in 100 runs of the same question.
- Measure a rate, not a rank: the useful number is the share of answers that name you across repeated runs, not your position in a single answer.
- Log the sources: the pages AI tools cite show which sites they trust for your topic, and where to earn mentions next.
- Repeat it monthly: keep the same questions, tools and number of runs so changes reflect real movement rather than noise.
Most people check their AI visibility once. They ask ChatGPT for the best option in their category, see who comes up, and either relax or panic. Both reactions are premature, because the same question asked again can return a different list.
What you really want to know is simple. How often does AI recommend you? Who gets named instead? And what would change that? A single screenshot can't answer any of those, and acting on one can send your budget after the wrong problem.
This guide shows how to run an AI visibility audit that gives you a number you can trust and track. It covers what to measure, how many times to ask, and how to score and read the results. It's based on published research on how AI recommendations vary, plus Google's own reporting tools.
What an AI visibility audit measures
An AI visibility audit measures how often AI tools name your brand when buyers ask about your category. It also shows who they name instead and which sources they rely on. Five measures cover it:
| Measure | What it tells you | Where it comes from |
|---|---|---|
| Mention rate | The share of answers that name you | Your own repeated test runs |
| Competitors named | Who AI recommends when it doesn't recommend you | The same runs |
| Sources cited | Which sites the AI tools trust for your topic | The links shown in each answer |
| Accuracy | Whether what AI says about you is true and current | Questions about your brand by name |
| Google AI impressions | How often your pages appear in Google's AI features | Google Search Console, where available |
Notice what's missing: a ranking position. There's a good reason for that.
Why one check tells you almost nothing
AI answers are built fresh each time, so the brands they name shift from one run to the next. In January 2026, SparkToro published research with the AI tracking company Gumshoe. Volunteers ran 12 recommendation questions 2,961 times across ChatGPT, Claude and Google's AI answers.
- Same list: ChatGPT and Google's AI returned the same list of brands less than once in 100 runs. Claude was only slightly more consistent.
- Same order: the same brands in the same order appeared roughly once in 1,000 runs.
- What held steady: how often a brand appeared across many runs. SparkToro's Rand Fishkin concluded that visibility percentage across repeated runs is "a reasonable metric", while ranking position in an answer isn't.
The research has limits. It was run with a company that sells AI tracking, and it isn't peer-reviewed. But the lesson for an audit is clear: ask each question several times and count how often you appear. Don't report where you ranked.
How to run the audit, step by step
Set aside a few hours and a spreadsheet. For 12 questions across three tools with three runs each, you'll review 108 answers.
- Write your buyer questions: 10 to 15 questions your buyers ask before they buy. Mix category questions ("best payroll software for a 50-person company"), problem questions ("how do I cut payroll errors") and comparisons. Add two about you by name, such as "What is [your brand] known for?", to check accuracy.
- Pick the tools your buyers use: usually ChatGPT, Perplexity and Google's AI Overviews or AI Mode. Use a private or logged-out window so your own history doesn't shape the answers.
- Run each question several times: three runs per question per tool, ideally on different days. Use the exact same wording each time.
- Log who gets named: for each answer, record whether you're named, which competitors are named, which pages are cited and any wrong facts about you.
- Score your mention rate: divide the answers that name you by the total answers. Do the same for each competitor so you can compare.
- Fix the biggest gap: use the patterns in the next section to choose one thing to fix first, then re-run the same audit next month.
Here's what part of a log might look like for an example payroll company:
| Question and tool | Named in (of 3 runs) | Cited sources | Notes |
|---|---|---|---|
| Best payroll software for 50 staff · ChatGPT | You 0 · Competitor A 3 · Competitor B 2 | A review site, Competitor A's pricing page | Not named |
| Best payroll software for 50 staff · Perplexity | You 1 · Competitor A 3 | A review site, a Reddit thread | Named once, third |
| What is Example Payroll known for? · ChatGPT | You 3 | Your About page | Lists a plan you dropped last year |
If you're named in 13 of 108 answers, your mention rate is 12%. The number matters less than the comparison: if Competitor A sits at 60% on the same questions, you know the size of the gap.
How to read your results
Your log will usually point to one of five patterns. Each has a different fix, so read the pattern before you spend anything.
| What you see | What it usually means | Fix first |
|---|---|---|
| Competitors named, you never are | AI tools don't see enough about you on other sites | Earn mentions on the sites being cited |
| Named on Perplexity, never on ChatGPT | One tool may not be able to reach your site | Check robots.txt allows OAI-SearchBot |
| Your pages cited, your brand not named | Your content is useful but not linked to a clear recommendation | Make your pages say plainly who you're for |
| Named, but with wrong facts | Old or conflicting information online | Update your site and business profiles to match |
| Named only when asked by name | AI knows you exist but doesn't link you to the category | Build content and mentions around category questions |
The "Cited sources" column does extra work here. A short list of sites will come up again and again for your topic. Those are the places worth a mention, and the places your competitors are already showing up. Our guide to getting your brand cited in AI answers covers how to earn them.
Free checkers, paid trackers or doing it yourself
You can run this audit by hand, with a free online checker or with a paid tracking tool. They suit different stages.
| Option | Good for | Watch out for |
|---|---|---|
| Free online checker | A quick first look in minutes | Few runs per question, so the score can swing a lot |
| Doing it yourself | Seeing the actual answers and wrong facts | Takes a few hours; hard to run at volume |
| Paid tracker | Many runs a month and a trend line | Ongoing cost; check how many runs sit behind each figure |
Whichever you use, remember Google's warning in its guide to generative AI search: "no third-party tool has access to our internal ranking or AI systems". Tools can measure what answers say. None can see why.
On our AEO service, this audit is where the work starts. We run a fixed list of buyer questions through the main AI tools every month and report the share of questions where you're named. If you're still deciding how AI search fits with your SEO, start with AEO vs SEO.
FAQs
How do I check if ChatGPT recommends my business?
Ask ChatGPT the questions your buyers ask before they buy, such as "best payroll software for a 50-person company", in a logged-out or private window. Run each question several times, because the brands named change from run to run. Count how many answers name you, note which competitors appear instead and which pages are cited. Repeat with Perplexity and Google's AI answers to see the full picture.
What is a good AI visibility score?
There's no industry standard yet, and tools calculate scores in different ways. The useful comparison is against your own competitors on the same questions. If a competitor is named in 40% of answers and you're named in 5%, that gap matters more than any single number. Track the same question list every month and judge yourself on the trend.
How often should I run an AI visibility audit?
Once a month is enough for most businesses. AI answers shift from day to day, so checking more often mostly shows noise rather than progress. Keep the question list, the tools and the number of runs the same each month, so changes in your mention rate reflect real movement. Re-run it sooner after a big change, such as a site migration or a major press mention.
Are free AI visibility checkers accurate?
They're useful for a quick first look, but treat the score with care. Some run each question only once, and research by SparkToro found ChatGPT and Google's AI give the same list of brands less than once in 100 runs. Google also says no third-party tool has access to its internal ranking or AI systems. Use a free checker to spot obvious gaps, then confirm by running the questions yourself.
Can I see AI Overviews data in Google Search Console?
Partly. Google counts clicks from AI Overviews and AI Mode in the normal Performance report, under the Web search type. In June 2026 it also began rolling out Search Generative AI performance reports to some sites. They show impressions in AI Overviews, AI Mode and Discover by page, country and device. At launch, they didn't show which searches triggered them.
Guy founded Wincon Marketing in Bangkok after nearly ten years in SEO, content and paid media, both agency-side and in-house. He moved early into AEO as AI assistants started answering buyers' questions, and now helps brands get found on Google and in AI answers.
Connect on LinkedIn →