How to Measure AI Search Visibility in Search Console in 2026

To measure AI search visibility in Search Console, use the Generative AI performance report for impressions in AI Overviews and AI Mode. Then pair it with the regular Performance report for clicks and with your analytics referral data. The AI report shows impressions only, with no clicks, positions or queries, so no single report tells the whole story.

For two years, SaaS content teams had no Google data on AI answers at all. AI Overview impressions sat inside the regular Performance report with nothing to separate them. That changed this summer, and the new report is now available to every site.

This guide is for SaaS marketers and founders who report on content performance to a boss, a board or themselves. It covers what the new report shows, what it leaves out and a monthly routine that turns three partial data sources into one honest picture. Cekpr covers measurement for search and AI visibility, and this is the reporting setup we recommend to teams starting from zero.

What does the Search Console Generative AI report show?

The report shows how often URLs from your site appeared in generative AI features, such as AI Overviews and AI Mode, plus generative AI features in Discover. You can break impressions down by page, country, device and date. Google announced the Generative AI performance reports on June 3, 2026.

The rollout started with a subset of sites. Google's announcement was later updated to say the reports reached all websites worldwide as of August 31, 2026. If you checked in July and saw nothing, look again.

Date granularity is unusually fine. The report supports hourly, daily, weekly and monthly views. Device data is available for Search, not Discover.

Hourly data is useful for one job: checking whether a new page shows up in AI answers within days of publishing. For trend reporting, stick to monthly totals so a single busy afternoon doesn't distort the picture.

What the Generative AI report leaves out

The gaps matter as much as the data. The Generative AI report has no clicks, no click-through rate, no average position and no query data. You can see that a page appeared in an AI answer, but not which question triggered it or whether anyone clicked.

Clicks aren't missing from Search Console entirely. Google's documentation on AI features and your website says traffic from AI Overviews and AI Mode is counted in the overall Performance report under the Web search type. It just isn't broken out from classic results.

One practical warning follows from that. The AI impressions already sit inside your Web totals. If you add the two reports together in a dashboard, you double count.

A monthly routine to measure AI search visibility

The routine below takes about 90 minutes a month once the sheet is set up. Each step adds a layer the previous one can't provide.

Step six is the one that keeps the routine alive. A spreadsheet nobody reads gets abandoned by the third month. A three-sentence summary gets forwarded.

Which metrics show AI search visibility for a SaaS site?

Four metrics cover most of what a SaaS team needs: AI impressions by page, Web clicks on those pages, AI assistant referral sessions and branded search demand. Each answers a different question, and each has a blind spot. Used together, they show reach, traffic and downstream interest.

The table below lays out where each metric comes from and what it can't tell you.

Branded demand is the sleeper metric. When AI answers cite you without sending a click, some buyers search for your brand later. A steady rise in branded queries alongside rising AI impressions is a reasonable sign the citations are working.

Why AI impressions rarely turn into clicks

Expect a low click rate from AI answers and plan your reporting around it. Pew Research Center studied 900 U.S. adults' Google searches in March 2025. Users clicked a traditional result in 8% of visits to pages with an AI summary, against 15% without one.

Clicks on the sources inside the summary were rarer still. Pew found them in about 1% of visits to pages with an AI summary. A page can earn thousands of AI impressions and only a handful of visits.

That's why the report needs context before it reaches a dashboard. Present AI impressions as reach, like a brand metric. Judge them against branded demand and pipeline, not against classic organic click-through rates.

Lessons from large-scale analytics programs

Big organizations face the same tracking problem at a larger scale. The federal government's Digital Analytics Program, run by the General Services Administration (GSA), measures thousands of government websites under one shared Google Analytics 360 account. The point of that setup is consistency: every site counts visits the same way.

The program also publishes its numbers. The public analytics.usa.gov dashboard shows federal web traffic over the last 30 days. A SaaS team doesn't need that scale, but it does need the same discipline: one definition per metric, written down and used every month.

Common mistakes in AI search visibility reports

Most broken AI reports fail in one of a few predictable ways. Each one makes the numbers look better or worse than they are. The list below covers the ones worth guarding against from the first month.

At Cekpr, we add one more rule: write the definitions on the first tab of the sheet. Which referrers count as AI, which pages are in scope, and which date range each figure covers. A new teammate should be able to rebuild the report from that tab alone.

Definitions also protect you when numbers drop. If a stakeholder asks why AI impressions fell, a documented method lets you separate a real change from a quirk in how the data was pulled.

How to raise AI visibility once you can measure it

Measurement only pays off when it changes what you publish. Pages that gain AI impressions usually answer a specific question with specific facts. Pages that lose them are often thin or outdated.

Start by making sure AI tools can reach your pages at all, then decide which extras are worth your time. Our explainer on whether a SaaS site needs an llms.txt file covers one popular extra and what Google says about it.

Accuracy matters too, because an AI answer repeats your claims at scale. Run new pages through the review steps in our guide to fact-checking AI-assisted content. And if traffic fell during the autumn rollout window, rule out link problems with our post-update backlink audit before blaming AI answers.

Start your AI visibility report this month

The answer is three sources, joined by page. Generative AI impressions show reach, Web clicks show traffic and analytics referrals show visits from other AI tools.

Export last month's data this week and build the joined sheet once. Next month, the routine takes 90 minutes, and you'll have the first trend line your team has ever had for AI search.