Visibility in AI search means being cited, mentioned or linked in an AI-written answer. It is harder to measure than a ranking because the answers change by question, place and run, and the engines publish few reports.
The method below is deliberately simple: a few official reports, one repeatable sample, referral data with its limits, and outcomes.
What can the engines tell you?
Google says that sites appearing in its AI features are included in overall search traffic in Search Console, reported in the Performance report under the Web search type. That means AI Overviews and AI Mode traffic is counted, but you cannot separate it from ordinary traffic in that report. Google's May 2025 post advises looking at the overall value of visits from Search, such as sales, sign-ups and information lookups about your business, and not at clicks alone.
Microsoft offers more. Bing Webmaster Tools has an AI Performance report, and Microsoft's 16 June 2026 post says it shows where your content is cited in AI-generated answers across Copilot, Bing and select partner experiences. New preview views group the grounding queries by intent and topic, show a citation share for a query, and compare periods. Microsoft says citation share is an observational metric, is not traffic share and does not expose competitor domains.
How do you sample AI answers fairly?
Answers differ by question wording, location, account history and model version, and the same question can give different answers on different runs. A single screenshot proves little. A fair sample is repeatable:
- Write a fixed list of 20 to 30 questions your buyers ask, in their words, covering problem, comparison and brand questions.
- Check them the same way each month: same tools, same location, signed out where possible, one person doing it.
- Record what you see: whether your business is mentioned, whether your page is linked, which other sources appear, and the date.
- Read the trend across runs. Look for movement over several months, and treat a single change as noise.
- Keep the list stable. Add questions rather than replacing them, so the history stays comparable.
What does referral data show, and miss?
Look in your analytics for visits whose source is an AI tool, such as chatgpt.com or perplexity.ai. These visits are real, but the count is partial: some tools open links without passing a referrer, so some visits arrive as direct traffic, and a visit tells you nothing about answers that never produced a click.
We could not find documentation from the engines that explains how they label outgoing links, and reports from analytics vendors on the subject disagree, so we give no numbers. Use referral data as a floor and a trend, never as the total.
Which outcomes should you count?
Count what the business wants: calls, form sign-ups, bookings, quotes and sales, with the source recorded. Ask every new customer how they found you, and give 'an AI assistant' as an answer option. Pew Research tracked 68,879 Google searches by 900 US adults in March 2025 and found that users clicked a result link in 8% of visits with an AI summary, against 15% without, so some of the value now arrives by name, phone or later search.
Branded search in Search Console is a useful side signal: a rise in people searching your name may reflect AI answers that mention you. Treat it as a hint only.
What about AI visibility tools?
Several SEO suites and specialist products sell AI visibility tracking, usually by running a set of questions through AI tools on a schedule and counting mentions. They automate the sampling above and can save time. Check how many questions they run, how they vary location and account, how they record sources, how they handle the run-to-run differences, and what they cost. Their numbers come from samples of answers and are separate from the engines' own data. We name no tool here and have no partner arrangement with any.
What does a simple monthly report look like?
- Engines: Search Console clicks and impressions for your key pages; Bing's citations and grounding queries for the same pages.
- Sample: the fixed question list, with mention, link and other sources for each.
- Referrals: visits from AI tools, with the caveat that the count is partial.
- Outcomes: calls, sign-ups and sales by source, and what new customers say about how they found you.
- Notes: pages changed this month and what you expect to see.
What should you avoid?
Avoid reporting a single answer as a result, judging by one tool's score, comparing across tools as if they were the same measure, and promising a client a position in an answer. Avoid changing the question list every month. And avoid spending on tracking before the pages themselves are clear, accurate and reachable. The pillar guide on how to optimize for AI search covers that work, and our guide on measuring digital marketing ROI shows how to tie it to revenue.
Conclusion
Tracking AI search visibility means combining a few official reports, one repeatable sample, partial referral data and real outcomes. None is complete alone, and together they show whether the work is helping.
PromotEdge International is an international strategy, technology, marketing and design company, and our delivery team is in India. See our tracking and data setup page and our AI search service.
Facts we rely on: see our company facts for what we state about ourselves and where each fact comes from.