How to measure and report AI search visibility in 2026

AI search visibility is how often your brand and your URLs show up inside answers from Google AI Overviews, Google AI Mode, ChatGPT, Perplexity, and Microsoft Copilot. This guide covers where the data lives in 2026, which four metrics to report, and how to pull Google Search Console, Google Analytics 4, and citation data into one report.

Your CMO forwards a screenshot. Someone asked ChatGPT which reporting tools their team should shortlist, and the answer names three of your competitors. Yours isn't in it.

So you go looking. Google Search Console shows impressions and clicks, both holding steady. GA4 shows organic sessions, roughly flat. But neither one tells you whether an assistant mentioned you last Tuesday, or which page it pulled from, or who it named instead.

AI search surfaces have stopped reporting like a results page, and the tools you've used for a decade were built to count positions and clicks, back when using AI in marketing meant something else entirely. Does this mean you can’t measure or report on AI search visibility? No, not at all. It just takes combining four numbers from four places, and a bit of discipline about how often you go and get them.

As the content strategist here at Supermetrics, I've recently been redesigning how we measure and report AI visibility. In this article, I'll explain what I've discovered and give you the exact steps I follow to measure and report our results with confidence.

Key takeaways

• Google Search Console's Search Generative AI performance reports, announced by Google in June 2026, report impressions for AI Overviews, AI Mode, and AI features in Discover, broken down by page, country, device, and date.
• The Search Console generative AI report carries no click, click-through rate, or query data, a limitation Search Engine Land flagged when the reports rolled out in June 2026.
• Bing Webmaster Tools launched its AI Performance report in public preview in February 2026 and added Citation Share in June 2026, which Microsoft defines as the percentage of citations attributed to your site out of all citations shown across all sites for that same grounding query (the prompt or search term Copilot uses to fetch results).
• Google Analytics 4 (GA4) assigns visits from AI assistants to an AI Assistant channel in the Default Channel Group and tags them with the medium "ai-assistant", according to Google Analytics Help.
• Ahrefs, analyzing December 2025 Search Console data across 300,000 keywords and publishing in February 2026, found that the presence of an AI Overview correlated with a 58% lower average click-through rate for the top-ranking page, up from the 34.5% Ahrefs measured in April 2025.
• Prompt-level presence rate, the fourth metric, only becomes reliable when you sample each prompt three to five times per engine, because the same prompt run once can return two different answers.
• Verdict: no single tool reports AI search visibility end to end in 2026, so combine impression data from Search Console and Bing Webmaster Tools, citation data from a GEO tracker, and referral data from GA4 into one report on a fixed monthly refresh.

What is AI search visibility, and how is it different from keyword rankings?

AI search visibility is how often your brand and your URLs appear inside answers generated by AI search surfaces, including Google AI Overviews, Google AI Mode, ChatGPT, Perplexity, Microsoft Copilot, and Gemini. Keyword rankings measure where your page sits on a results page. AI search visibility measures whether your content made it into the answer, which often happens without anyone clicking.

Two things get counted in AI search visibility, and they're worth separating. A mention is the assistant naming your brand in its answer with no link attached. A citation is the assistant linking your specific URL as a source. Assistants name brands more often than they link them, so a citation-only view will tell you that you're invisible on a surface where you're being recommended.

The click layer underneath both mentions and citations keeps thinning. Pew Research Center, analyzing the March 2025 browsing behavior of 900 US adults in a report published in July 2025, found that visits to pages with an AI summary ended in a click on a traditional search result 8% of the time, compared with 15% on pages without one.

Rankings feel the same squeeze. Ahrefs, comparing 150,000 keywords with an AI Overview against 150,000 informational keywords without one using December 2025 Search Console data, found the top-ranking page took a 58% lower average click-through rate when an AI Overview was present. In April 2025, the same study put that gap at 34.5%.

In other words, your page can still hold position one in the search engine results pages (SERPs) and lose the traffic anyway. That's why the reporting has to change, and why AI visibility is now a crucial figure to include in reporting, sitting alongside the classic metrics from across your marketing reporting stack.

Which AI search visibility metrics should you report on?

Report four: AI impressions, citation share, AI referral sessions, and prompt-level presence rate. Each one comes from a different source, and no source gives you more than one of them.

1. AI impressions count how often your pages appeared inside AI features. Google Search Console supplies this for AI Overviews and AI Mode. Bing Webmaster Tools supplies the equivalent for Microsoft Copilot.
2. Citation share tells you how much of the available citation space you're winning, on any assistant. Microsoft's own Citation Share metric, capitalized, is the Bing-specific version of this idea, per the Bing Webmaster Tools update of June 16, 2026.
3. AI referral sessions count the people who arrived. GA4 handles this one, and it's the only metric of the four that connects to revenue.
4. Prompt-level presence rate tells you how often you're named when a real buyer question gets asked. You measure it by running a fixed set of prompts across each assistant and recording how often your brand appears. Ask the same prompt twice and you'll often get two different answers, so a single run proves nothing. Sample each prompt three to five times per engine in a logged-out session, then record the rate rather than the outcome.

None of this replaces your organic baseline. Keep the classic impressions, clicks, and position data running alongside it, using the same SEO analytics reporting process you already have, because AI Overviews are built on the same index that decides those rankings.

Can you see AI Overviews data in Google Search Console?

Yes, partly. Google announced Search Generative AI performance reports in Search Console in June 2026, and they report impressions for AI Overviews, AI Mode, and AI features in Discover, broken down by page, country, device, and date. They don't report clicks, click-through rate, or queries.

The missing click data shapes what you can say. The report answers "did our pages appear in Google's AI features, and which ones." It can't answer "what did that traffic do," which is the question your CMO will ask next. Google doesn't tell site owners how many searchers click through from an AI response.

Search Console carries a second constraint that catches people out later: it only keeps 16 months of performance data on a rolling window, so anything older just disappears. AI impression data only starts accumulating in 2026, which means the first year-over-year comparison you'll ever be able to run depends entirely on whether you exported the data before the window closed.

If you want to keep all your Search Console data in one place, then you need to use Supermetrics, the marketing intelligence platform that connects data from 150+ marketing and sales platforms into the reporting tools, warehouses, and AI assistants marketing teams already use.

Supermetrics' Google Search Console connector pulls that performance data into Google Sheets, Excel, Looker Studio, or Power BI on a schedule you set. Supermetrics then keeps a governed historical record of this data, so Search Console's 16-month window stops being the limit on how far back your trend line goes. Plans include a free trial, so you can test the Search Console pull before committing to the history.

How do you track AI referral traffic in Google Analytics 4 (GA4)?

Use the AI Assistant channel in GA4's Default Channel Group, then add a custom channel group to catch what it misses. Google Analytics 4 assigns detected visits from AI assistants to an AI Assistant channel, tags them with the medium "ai-assistant", and labels the campaign "(ai-assistant)", according to Google Analytics Help.

To find the AI Assistant channel, open Reports, then Acquisition, then Traffic acquisition, and set the primary dimension to Session default channel group. For anything the native channel doesn't recognize, create a custom channel group with a Session source regex covering chatgpt\.com|chat\.openai\.com|gemini\.google\.com|claude\.ai|perplexity\.ai|copilot\.microsoft\.com, and order it above Referral so it evaluates first. If you need a refresher on how to use the tool, our Google Analytics 4 reporting guide covers the underlying setup.

One warning before you present the AI referral number. Some assistants send visits without a referrer header, and GA4 files those under Direct like any other unattributed visit. Your AI referral count is a floor. Report it as a minimum and say so out loud, rather than letting someone else discover it.

Counting sessions is the easy half. Segment those sessions by conversion, sign-up, or pipeline, because a small volume of AI referrals converting at three times your organic rate is a very different story from the same volume bouncing. The Supermetrics Google Analytics 4 connector pulls both the sessions and the conversions into the same destination as your Search Console data, which is what makes the comparison possible without a manual export.

How do you track citations in Copilot, ChatGPT, and Perplexity?

Microsoft reports Copilot citations directly to site owners through Bing Webmaster Tools. OpenAI and Perplexity publish no equivalent report, so tracking ChatGPT and Perplexity means a third-party GEO (generative engine optimization) tracker or a prompt panel you run yourself. The two methods don't produce equally reliable numbers, so label which is which when you report them.

Tracking Copilot citations in Bing Webmaster Tools

Among the assistants covered in this guide, only Microsoft Copilot reports citation data directly to site owners. Bing Webmaster Tools launched AI Performance in public preview in February 2026, showing how often your URLs are cited across Microsoft Copilot, AI-generated summaries in Bing, and selected partner integrations.

Microsoft expanded the AI Performance report on June 16, 2026 with four additions: Intents, which classify the grounding queries Copilot generates internally; Topics, which cluster related queries; Citation Share; and Compare, which overlays a previous period. Microsoft is explicit about the boundaries, describing Citation Share as an observational metric that doesn't represent traffic share, doesn't expose competitor domains, and doesn't assign a quality score. Treat it as a direction of travel.

Sampling prompts for ChatGPT and Perplexity

Building a prompt panel is the direct way to measure ChatGPT and Perplexity citations yourself, without waiting on either company to publish a report.

Build the prompt panel from questions people ask: your Search Console queries, People Also Ask phrasing, and the questions your sales team hears on calls. Run each prompt three to five times per engine in a clean, logged-out session, and record how often your brand appears and which competitors appear alongside it. A screenshot of one run is an anecdote.

If you'd rather have the citation data land in your reporting stack automatically, Adthena, the search intelligence platform, has an AEO (answer engine optimization) and GEO connector in Early Access with Supermetrics. It tracks how often AI Overviews appear by search term split across desktop and mobile, plus ChatGPT and Perplexity citation frequency and average position, and delivers it to Google Sheets, Excel, Power BI, Supermetrics Data Explorer (our natural-language interface for querying connected data), a data warehouse, or the API.

Treat any third-party GEO tracker's numbers as directional. Sampling methodology varies by vendor, and most don't publish it, so a citation count from one tracker isn't safely comparable to a count from another. Pick one tracker and stay with it for the full reporting period rather than switching mid-quarter.

How do you build an AI search visibility report your CMO will read?

Put all four metrics on one page, one row per surface, each with a trend line, refreshed monthly.

The four numbers usually stall in four different tools, with nobody owning the job of joining them. Combining them follows a fixed sequence:

1. Pull AI impressions from Google Search Console and Bing Webmaster Tools.
2. Pull AI Assistant sessions and their conversions from GA4.
3. Pull citation frequency and position from your GEO tracker.
4. Record your prompt panel presence rate for the month.
5. Standardize how each surface is named before any of it goes into a chart.

Standardizing surface names is the step people skip. Search Console, GA4, and a citation tracker will each label the same surface differently, so "ChatGPT" arrives as three values and your chart splits one surface into three. Supermetrics data blending, a feature that merges rows from multiple connected sources into one queryable table, solves the first half of that problem. Custom Fields, which let you standardize a value like a surface name across every source at once, solve the second: define "ChatGPT" once and it means the same thing in every row.

Then decide where the report lives. A hand-built spreadsheet has a predictable half-life: someone rebuilds it once, then it drifts until nobody trusts the numbers, and eventually it gets abandoned. Supermetrics Studio hosts the dashboard on live data instead, so your CMO (or your clients, if you're reporting for a marketing agency) can check the trend without asking you to refresh anything.

If you're running AI visibility reporting for clients rather than one brand, split it with Workspaces. Each client gets its own workspace with its own connected accounts and its own access rules, so a client only ever sees their own data, and your surface definitions stay consistent across all of them. The same structure serves a single enterprise running multiple brands or regions: each one gets its own workspace, standardized on the same metric definitions, instead of every regional team rebuilding its own version of the report.

If you don't have the budget for a dedicated GEO tracker yet, start with what's free. Search Console and GA4 cover two of the four metrics at no extra cost; add a citation tracker or a manual prompt panel once that baseline is running on a steady monthly cadence.

What can't you measure about AI search visibility yet?

You can't see clicks from Google's AI features, the prompt behind a ChatGPT or Perplexity mention, and you can't get a complete count of AI-driven sessions. Gemini shares the click gap too: it draws on the same Google index as AI Overviews and has no separate citation report of its own, so track it as part of your Google numbers rather than as a fifth data source.

There's also a fourth gap: an assistant recommendation often converts weeks later as a brand search or a direct visit, and no attribution model will connect the two. Track brand search volume and direct traffic alongside your AI referrals so the influence at least shows up somewhere.

And resist the temptation to ask an AI assistant to estimate your own visibility, because an assistant without access to your data will produce a confident number with nothing behind it, which is why ungrounded AI agents hallucinate marketing numbers more often than their confident tone suggests.

Supermetrics can’t help you with prompt sampling, share of voice inside ChatGPT or Perplexity answers, or modifying your content to increase AI search visibility. However, what it can help with is to connect the data that already exists, from Search Console, GA4, and a GEO tracker like Adthena, into one place you can report from (using Supermetrics Studio to keep reports up to date and trustworthy).

How often should you report on AI search visibility?

Report monthly, sample your prompt panel weekly while a change is in flight, and re-baseline after any major engine update. Monthly matches the pace at which retrieval indexes move, so a weekly report tracks index noise more than real movement.

Timing matters when you ship a change. Recrawling and re-embedding both lag, so our content team at Supermetrics rechecks citations at around four weeks and organic positions at six to eight weeks. Reading the numbers earlier than that tells you very little.

Score the result against the competitors cited for each prompt rather than against an absolute target. Going from unmentioned to co-cited with two rivals is a real win, even when your citation rate barely moves.

One last thing about how the report lands. Supermetrics' 2026 Marketing Data Report found that only 18% of the 435 marketers surveyed across the US, UK, Germany, Australia, and Singapore report high trust in AI. In a room where trust in AI-related numbers is that low, a report showing four sourced metrics and their limitations will travel further than one showing a single composite visibility score nobody can trace.

What to tell your CMO about AI search visibility

You can tell them four things: how often your pages appeared in Google's AI features last month, how much of the Copilot citation space you held, how many sessions arrived from assistants and what they did next, and how often you were named across a fixed panel of buyer questions. When your CMO forwards that screenshot again, "I don't know" stops being the only available answer.

None of those four numbers is complete on its own, and the honest report says so. Together they show direction, which is what the decision needs.

Start with the two sources you already own. Connect Google Search Console and Google Analytics 4, get the impressions and the referral sessions into one place, then add citation data on top. Finally, report it all in Supermetrics Studio, creating customizable dashboards that stay up to date with your AI search visibility data.

Start a free Supermetrics trial and build the first version of the report this week.

FAQs