Your customers now ask ChatGPT, Gemini, Perplexity and Google’s AI Mode which supplier to shortlist, and many of them never click a blue link. So the question every marketing head is asking is simple: are we being mentioned, and how would we even know?
AI visibility tracking is the process of measuring how often your brand appears, gets cited and gets recommended inside AI-generated answers for the questions your buyers ask. It combines three data sources: repeated prompt testing across AI platforms, first-party reports from Google Search Console and Bing Webmaster Tools, and AI referral traffic in Google Analytics 4.
This guide explains which metrics are worth tracking, which ones mislead, and how to build a monthly reporting process that your leadership team will trust.
Why AI visibility needs its own measurement
Traditional SEO reporting is built on rankings, impressions and clicks. AI answers break that model in three ways.
First, there is no stable “position”. An AI assistant writes a fresh answer every time, often naming different brands in a different order.
Second, influence often happens without a click. A buyer can read a recommendation in ChatGPT, then search your brand name directly a week later. That visit shows up as branded search or direct traffic, not as AI traffic.
Third, the audience is large enough to matter. OpenAI said in February 2026 that ChatGPT had around 900 million weekly active users. Even a small share of category research moving to AI assistants changes how buyers form shortlists.
This is why AI visibility tracking sits alongside SEO reporting rather than replacing it.
The metrics that matter (and the one that doesn’t)
Not every number an AI visibility tool shows you is meaningful. Here is how the common metrics compare.
| Metric | What it measures | How reliable it is |
|---|---|---|
| Mention rate | % of responses to a prompt set that name your brand | Reliable when measured across many prompts and repeated runs |
| Citation rate | % of responses that link to or cite your website as a source | Reliable; also directly actionable for content teams |
| AI share of voice | Your mentions as a share of all brand mentions in your category’s prompts | Useful for competitor benchmarking |
| Sentiment and framing | Whether the AI describes you accurately and positively | Needs human review; useful for spotting errors |
| AI referral sessions and conversions | Visits and leads from AI platforms in GA4 | Reliable but undercounted |
| Average rank position | Where your brand appears in an AI-generated list | Weak; changes on almost every run |
The case against rank position comes from research. In early 2026, SparkToro’s Rand Fishkin and Gumshoe.ai’s Patrick O’Donnell had around 600 volunteers run 12 identical prompts nearly 3,000 times across ChatGPT, Claude and Google’s AI. The chance of getting the same list of brands twice was under 1 in 100, and the same list in the same order was roughly 1 in 1,000.
What did hold up was visibility percentage. Some brands appeared in 60–90% of responses for a given intent, even when the exact list changed. That is the number to report.
How to track AI visibility: a six-step process
Step 1: Build a prompt set that reflects real buying intent
Start with 30 to 100 prompts grouped by intent, not by keyword. A useful mix includes:
- Category discovery: “best SEO agencies for SaaS companies in India”
- Problem-led: “how do I reduce cost per lead on Google Ads”
- Comparison: “[your brand] vs [competitor]”
- Brand checks: “what does [your brand] do” and “is [your brand] reliable”
- Location prompts where relevant: “digital marketing agency in Noida for healthcare”
Pull these from sales call notes, Search Console queries, People Also Ask results and the questions your account managers hear every week. Keep the set fixed for at least a quarter so trends are comparable.
Step 2: Choose the platforms your buyers use
For most B2B and services businesses, the core list is ChatGPT, Google AI Overviews and AI Mode, Gemini, Perplexity and Microsoft Copilot. Add Claude if your audience is technical. Track each platform separately, because they draw on different sources and can describe your brand very differently.
Step 3: Run each prompt more than once
Because answers vary, a single run tells you very little. Run each prompt several times per period and record whether your brand was mentioned, whether your site was cited, which competitors appeared and how you were described.
You can do this manually for a small prompt set in a spreadsheet. Beyond 30 to 40 prompts, a dedicated tracker becomes worth paying for. Tools in this space include Semrush, Ahrefs, SE Ranking and specialist platforms such as Profound and Peec AI. Features, platform coverage and pricing change quickly, so compare them against your own prompt set before committing.
Step 4: Use the free first-party reports
Three platform owners now give you direct data, and they cost nothing.
Google Search Console. AI Overviews and AI Mode impressions are counted inside the standard Performance report. In June 2026 Google also introduced separate Search Generative AI performance reports showing impressions, pages, countries, devices and dates for AI features in Search and Discover. These are rolling out gradually, so not every property has them yet, and they do not show clicks.
Bing Webmaster Tools. The AI Performance report, released in public preview in February 2026, shows total citations, cited pages and sample “grounding queries” across Microsoft Copilot, Bing’s AI-generated summaries and select partner integrations. It is one of the few places you can see which of your pages an AI system actually cited.
Google Analytics 4. Since May 2026, GA4 has a native AI Assistant channel in the Default Channel Group. Sessions from recognised assistants such as ChatGPT, Gemini and Claude get the medium “ai-assistant”. ChatGPT also adds a utm_source=chatgpt parameter to many outbound links, which helps attribution.
Step 5: Account for traffic you cannot see
AI referral numbers are a floor, not a total. Visits from AI mobile apps, in-app browsers and copied links often arrive without a referrer and land in Direct.
To get a fuller picture, keep a custom channel group for AI platforms that GA4 may not yet recognise, such as Perplexity and Copilot. Watch branded search impressions in Search Console alongside your AI mention rate. Add “How did you hear about us?” as a free-text field on lead forms, and ask your sales team to log AI mentions during discovery calls.
Step 6: Report monthly, act quarterly
AI answers are noisy week to week. A monthly report with quarterly decisions works well for most businesses. Each report should answer four questions:
- Did our mention rate and citation rate change for priority prompts?
- Which competitors gained or lost share of voice?
- Which of our pages are being cited, and which prompts cite competitors instead?
- Did AI referral traffic and AI-influenced leads change?
Turning tracking data into action
Tracking only pays off when it changes what you publish. The most useful pattern to look for is a prompt where competitors are cited and you are not. That gap usually points to one of these causes:
- No page directly answers the question. Create one with a clear answer near the top.
- Your page exists but is thin or outdated. Update it with specific facts, comparisons and current data.
- Third-party sources favour competitors. Review sites, industry directories, news coverage and community discussions often shape AI answers. Earning mentions there is part of the work.
- The AI describes you inaccurately. Make sure your About, service and location information is consistent across your website, Google Business Profile, LinkedIn and directories.
Common mistakes in AI visibility tracking
- Reporting a single screenshot. One answer from one session proves nothing either way.
- Treating AI rank as a KPI. Report visibility percentage instead.
- Tracking only branded prompts. You will look visible while missing the category prompts where buyers form shortlists.
- Changing the prompt set every month. Trends become impossible to read.
- Ignoring accuracy. Being mentioned with the wrong services or location can cost you leads.
Frequently asked questions
What is AI visibility?
AI visibility is how often and how accurately your brand appears in answers from AI platforms such as ChatGPT, Gemini, Perplexity, Copilot and Google’s AI Overviews. It covers whether you are mentioned, whether your website is cited as a source, and how you are described compared with competitors.
Can I track ChatGPT brand mentions for free?
Yes, at a small scale. Run a fixed set of prompts in ChatGPT several times a month and log the results in a spreadsheet. Pair this with GA4’s AI Assistant channel to see referral visits. Manual tracking becomes impractical beyond a few dozen prompts, which is when a paid tracker becomes worthwhile.
Does Google Search Console show AI Overviews data?
AI Overviews and AI Mode impressions are included in the standard Performance report. Google also launched separate Search Generative AI performance reports in June 2026, showing impressions, pages, countries and devices. They are being rolled out gradually and currently do not include click data.
How many prompts should I track?
Most businesses can start with 30 to 50 prompts covering category, problem, comparison, brand and location intents. Larger brands with several product lines may need 100 or more. Quality matters more than volume: each prompt should reflect a question a real buyer would ask.
How often should I check AI visibility?
Collect data weekly or fortnightly if you can, but report monthly and make strategic decisions quarterly. AI answers vary from run to run, so short-term movements are often noise rather than real change.
Why is my AI traffic so low in GA4?
Much AI-driven traffic arrives without referrer data, especially from mobile apps and in-app browsers, so it is recorded as Direct. Some buyers also read an AI answer and later search your brand name instead of clicking. Treat GA4’s AI numbers as a minimum and look at branded search trends as well.
Is AI visibility tracking different from GEO?
Yes. Generative engine optimisation (GEO) is the work of improving how AI systems understand and reference your brand. AI visibility tracking is the measurement that shows whether that work is succeeding. You need tracking to set a baseline and to prove return on GEO investment.
Final thoughts
AI visibility tracking is still a young discipline, and no tool gives a perfect picture. That is not a reason to wait. A fixed prompt set, repeated runs, the free reports in Search Console, Bing Webmaster Tools and GA4, and a disciplined monthly review will tell you far more than guesswork or one-off screenshots.
Focus on mention rate, citation rate and share of voice, not rank. Then use the gaps to decide what to publish and where to earn mentions.
If you want a baseline of where your brand stands today, eDigital Interactive’s AI SEO services team can audit your visibility across Google and AI platforms and show you where competitors are being cited instead of you.
