Google’s Search Console AI Reports Arrive Without Click Data

Google has started testing dedicated Google Search Console AI Search reports, an addition to the company’s free tool for tracking how sites turn up in Google. The reports show how often a site’s pages appear in AI Overviews and AI Mode, and a separate toggle lets owners pull their site out of those AI features entirely. Both are rolling out to a slice of UK websites first, with a global expansion promised after testing.

There is a catch that will frustrate the people who lobbied hardest for this. The reports count impressions, not clicks. After two years of asking Google to separate AI-driven visibility from ordinary search, site owners have been handed the visibility half and not the traffic half.

What Google Put in the New Reports, and What It Held Back

The new view is built around a single metric: impressions, meaning how often URLs from a site showed up inside generative AI features across Search and Discover. That data was already folded into the overall performance report; what is new is a place to look at it on its own, without AI activity blurring into the rest of a site’s search numbers.

Google says the dedicated reports break the impression data down several ways:

• Pages, so owners can see which specific URLs surfaced in AI features
• Countries, showing where the visibility came from
• Devices, for Search results
• Dates, with granularity down to the hourly level alongside daily, weekly and monthly views

The hourly detail is genuinely useful for anyone tracking how a page enters or drops out of an AI answer over a single day. What the reports leave out is the part that pays the bills. There is no click data and no query-level breakdown, so a site owner can see that a page appeared in an AI Overview but not whether anyone clicked through, and not which search prompted the appearance. Google laid out the structure in its Search Central announcement of the generative AI performance reports, and said it is still working out which further metrics to add.

Two Years of Asking for AI Data

To understand why this lands as half a win, rewind to May 2024, when AI Overviews launched in the United States after a year of testing under the Search Generative Experience banner. Once AI answers started occupying the top of result pages, the question for every publisher became simple: how much of my traffic is now flowing through these summaries?

Search Console could not answer it. When Google confirmed that AI Mode data counted toward Search Console totals, there was no way to pull AI-driven numbers apart from regular organic search. Everything sat in one bucket.

The measurement got murkier inside the AI answers themselves. John Mueller, a Google Search Advocate, explained that every link inside an AI Overview shares a single position in Search Console reports.

It can be first position, if the block is shown first.

That was Mueller describing how the whole AI Overview counts as one block, which means a site buried fourth or fifth among the cited links reads the same as one shown first. For more than a year, that opacity has been the recurring complaint from the SEO community, and the new reports answer one part of it while leaving the position-and-click puzzle intact.

Does Opting Out of AI Search Hurt Your Rankings?

The second feature in the test is the one that will make publishers stop and read the fine print. It is a toggle that lets a site exclude itself from AI Overviews, AI Mode and AI Overviews in Discover. Flip it, and the site stops receiving any traffic or impressions from those surfaces.

The reassurance Google offers is direct: the control will not be used as a ranking signal for search results outside of those AI features. In plain terms, opting out of the AI summaries should not push a site down in the regular blue-link results or remove it from the standard Discover feed. The site simply stops appearing inside the AI-generated answers.

How It Differs From Google-Extended

Google frames the toggle as the next step after two existing controls, and the distinction matters. Snippet controls govern how content is displayed in traditional search results. The separate Google-Extended crawler control documented for AI training lets a site block its content from being used to train Gemini and related models, without touching how Googlebot indexes the site for search. Neither of those decides whether a page shows up live inside an AI Overview.

What the Toggle Cannot Promise

The new toggle fills that last gap, covering appearance in the live AI experience rather than training or display. The trade is blunt for anyone who relies on AI-driven referrals: turning it on means giving up whatever visibility, and whatever clicks, those features currently send. With no click data in the reports, owners are being asked to make that call without knowing what they would actually be forfeiting.

Bing Got Here First

Google is not setting the pace here. Microsoft moved earlier and faster on AI search reporting through Bing Webmaster Tools, its own free measurement suite. The AI Performance dashboard arrived in public preview on February 10, 2026, letting publishers see how often their content is cited across Microsoft Copilot, Bing’s AI summaries and select partner integrations.

Microsoft pitched it plainly when the preview went live.

For the first time, you can understand how often your content is cited in generative answers, with clear visibility into which URLs are referenced and how citation activity changes.

That was the message accompanying the Bing Webmaster Tools public preview. The dashboard added grounding query-to-page mapping in March, surfacing the phrases that triggered a citation, and at SEO Week this spring, Krishna Madhavan, a principal product manager at Microsoft AI and Bing, previewed a metric called Citation Share that would show what percentage of citations a site captures within a given query. None of those Bing previews carry release dates yet, but the direction is set.

Neither tool hands over clicks, a reminder that AI answers do not generate referrals the way ten blue links did. On the other rows, Bing currently offers the deeper view, with query-level grounding data Google has not matched.

Impressions Without Clicks Tell Half the Story

Here is why the missing column matters so much. An impression confirms a page appeared in an AI answer. It says nothing about what happened next.

AI Overviews are designed to resolve a query on the spot, so a high impression count can sit alongside flat or falling visits, the so-called zero-click result. Without click data, an owner watching impressions climb cannot tell whether that exposure is sending readers to the site or quietly replacing the visit they used to get. The two readings point in opposite directions, and the report shows only the first.

What the new reports do enable is real, if narrow. A site can now confirm which pages are being pulled into AI answers, watch that presence shift by country and device, and react within hours rather than guessing from a bundled total. For diagnosing whether a content change moved a page in or out of AI surfaces, that is a step Search Console could not take before. It just stops short of the number that decides whether any of it is worth the effort, and Google’s own documentation for the Search Console performance report still treats clicks as the headline metric everywhere else in the tool.

When It Reaches Everyone Else

For now the test sits with a subset of UK site owners, and most publishers cannot touch either feature yet. Google says it will expand availability globally after the trial, without naming a date.

The company also says it is continuing to work with website owners to understand which insights will be most helpful, and that it plans to add more metrics to the reports over time. It has not named which metrics or when, which leaves the click-data question exactly where it has sat since AI Overviews launched. The first dedicated window into AI visibility is real, and the most important pane in it is still missing.