Summary
AI visibility is about how often your brand appears in AI answers, not where it ranks. Test the questions real buyers ask, run each one several times on every AI tool, and track how often you are mentioned, recommended and linked as a source. Check that the answers describe you correctly. Use the free AI data in Google Search Console, Bing Webmaster Tools and GA4, and keep track of the websites AI tools use as sources. Review the results monthly and look for trends. Sound measurement turns AI search optimisation from guesswork into a strategy.
Most brands check their AI visibility the same way. Someone types the company name into ChatGPT, reads the answer and reports back. It feels like you have the answer, but one answer is not a pattern. Ask the same question again and a different set of brands may appear, so one screenshot tells you little about the answer the next buyer gets.
The gap is wide. Semrush found 45% of marketing leaders cannot accurately measure their brand’s visibility in AI-generated answers, and only 9% have tools that track every relevant metric across platforms. This guide covers what to measure, where the free data sits, and how to turn it into one monthly number.
What Is AI Visibility?
AI visibility measures how often your brand shows up when AI tools answer questions in your category. Showing up can mean a mention, a recommendation, or a link to your site, whether in ChatGPT, Gemini, Perplexity, Copilot or Google’s AI Overviews.
It is how you measure the results of generative engine optimisation and answer engine optimisation. A Google ranking shows where your page appears in a list of links. AI visibility shows whether your brand is part of the answer, and how it is described.
Why One Prompt Check Is Not a Measurement
AI answers are probabilistic. SparkToro and Gumshoe ran brand-recommendation prompts repeatedly through ChatGPT, Claude and Google’s AI. Getting an identical list of brands twice happened less than once in 100 tries. Getting the same brands in the same order was rarer still, closer to once in 1,000.
The engines also disagree with each other. Semrush’s index of 126 million US prompts found that only 36 brands held top-100 visibility on ChatGPT, Gemini, AI Mode and AI Overviews in every month it studied.
Two rules follow:
- Run each prompt several times and record how often you appear, not where you ranked once.
- Report every engine on its own line. A blended score hides which engine you are missing.
Teams asking how to rank on ChatGPT tend to skip this step. You cannot improve what you have not measured.
The Metrics That Make Up an AI Visibility Score
A framework published in Search Engine Journal, sponsored by HubSpot, reduces the audit to three ratios, calculated per engine:
- Mention rate: prompts where your brand appears, divided by prompts tested.
- Recommendation rate: prompts where you are among the first three brands named, divided by prompts tested.
- Share of voice: your mentions, divided by all brand mentions across the answers.
Add two more:
- Citation rate: how often your own pages are linked as sources. An AI tool can name your brand but link to a review site instead. On Gemini, Semrush found this gap is wide, with named brands and linked domains overlapping as little as 30%.
- Accuracy: whether the answer describes your pricing, features and positioning correctly.
Build a Prompt Set That Mirrors Real Buyers
The prompt list decides whether the score means anything.
- Start with what already works. Use Google Search Console to find searches your site ranks for on page one, and keep the ones buyers use to compare options like “best”, “vs”, “alternative” or “pricing”.
- Rewrite each as the full question a buyer would type, with real constraints such as team size, city or budget.
- Run every prompt two or three times per engine in a logged-out or temporary chat, so saved memory does not influence the answer.
- Log the engine, the brands named and their order, whether you appear, and the domains cited.
A spreadsheet handles 25 to 50 prompts. Beyond that, AI SEO trackers such as HubSpot AEO do the same checks automatically at set intervals.
What the Platforms Now Report Directly
Google and Microsoft now share some of this data for free, though Search Console and Bing each report only on their own AI features.
Google Search Console
Google’s generative AI performance report shows how often your URLs appeared in AI Overviews and AI Mode, by page, country, device and date. It reached all sites worldwide on 31 August 2026 and, for now, reports impressions without clicks. It is a useful first stop for anyone figuring out how to appear in Google AI Overviews, because it shows what is already working.
Pair it with the branded queries filter, which separates searches for your name from everything else. A rise in these searches can hint that people first heard about you in an AI answer.
Bing Webmaster Tools
Microsoft’s AI Performance report counts citations across Copilot, AI summaries in Bing and selected partners, and lists the grounding queries behind them. A June 2026 update added Citation Share, which shows your site’s percentage of all citations for a given grounding query, and Compare, which sets the current period against an earlier one.
Google Analytics 4
ChatGPT adds utm_source=chatgpt.com to the links it sends, and GA4 now files recognised chatbot visits under a new AI Assistant channel. Clicks only tell part of the story, though, since many people read the answer and never visit your site. Previsible studied 6.77 million visits from AI chatbots and found that more than 9 in 10 came from ChatGPT. When ChatGPT sent fewer visitors in November 2025, total chatbot traffic across the sites studied fell by half in a single month.
Track the Sources That Shape the Answer
AI tools repeat what the web says about you. In Ahrefs’ study of 75,000 brands, branded web mentions correlated with AI Overview visibility at 0.664, against 0.218 for backlinks. The study shows a pattern rather than proof, but it suggests being talked about across the web matters more than collecting links.
Track which websites the answers rely on, then look at:
- Review sites, forums and roundup articles that recur across your prompts.
- Every unlinked brand mention on a site the engines already cite.
- Outdated or wrong claims about your pricing or product.
Keep your online reputation monitoring and brand reputation tool running for news and social media, and add your AI prompt checks to the routine.
What a Monthly AI Visibility Scorecard Looks Like
Review these once a month, engine by engine.
| Metric | Where to find it |
|---|---|
| Mention and recommendation rate, per engine | Prompt log or tracker |
| Share of voice against named competitors | Prompt log or tracker |
| Cited domains and your citation rate | Prompt log, Bing AI Performance |
| AI Overview and AI Mode impressions by page | Search Console generative AI report |
| Copilot citations and grounding queries | Bing Webmaster Tools |
| AI Assistant sessions and conversions | GA4 acquisition reports |
| Branded search demand | Search Console branded filter |
| Accuracy of pricing and feature claims | Manual review of logged answers |
Frequently Asked Questions
Start with 25 to 50 real customer questions. Ask each one a few times on ChatGPT, Gemini and Perplexity in a temporary chat, and record whether you’re mentioned, how high up, and which sites are linked. Support this with Search Console’s generative AI report, Bing’s AI Performance report and GA4’s AI Assistant channel, all of which are free.
Partly. The generative AI performance report shows AI Overviews and AI Mode impressions by page, country, device and date. It does not yet show clicks, and it covers only Google.
Check once a month and run each question a few times each time. SparkToro’s research shows AI answers change too often for one check to be reliable. The trend over several months will tell you more.
A mention is your brand named in the answer. A citation is a page the engine links as a source. Semrush found the two overlap as little as 30% on Gemini, so track both.
Start with the sources the engines already use for your prompts. Ahrefs found branded web mentions tracked AI Overview visibility far more closely than backlinks. Alongside that, apply AI content optimisation so AI tools can easily pull answers from your pages. ChatGPT SEO and LLM optimisation work best as steady, long-term efforts that you track over time.
