Search Console’s Generative AI Report Is Now Live for Every Site: What Your Practice Can See

As of August 11, 2026, Google finished rolling out the generative AI performance report to every Search Console property. The report first appeared in June and reached sites gradually over the following weeks, which meant some practices could open it while others still could not. That gap has now closed. Any practice with a verified Search Console property may access it today.

The report gives site owners a dedicated view of how their pages appear inside Google’s generative AI features. That includes AI Overviews and AI Mode in Search, along with generative AI features in Discover. Until this report existed, those impressions were folded into overall performance data with no way to separate them from traditional search results.

What the Report Shows

The report tracks impressions inside generative AI features and breaks them out in several ways. Practices may see how those impressions change over time, which pages earn the highest and lowest impression counts, and where the impressions originate by device and by country. For a practice trying to understand whether its service pages or its blog content is feeding AI-generated answers, the page-level breakdown is the most useful part of the report.

That visibility answers a question many practices have been guessing at for more than a year. AI Overviews have been reshaping how patients encounter health information, but until now there was no way to confirm which of your own pages were being drawn into those answers. The report replaces inference with something closer to measurement.

What It Does Not Show Yet

The report currently provides impressions only. There is no click data, which means a practice may learn that a page appeared inside an AI-generated answer without learning whether anyone actually visited the site afterward. This is a genuine limitation rather than a rollout delay, and it shapes how much weight the numbers deserve.

It also creates an interpretation problem worth naming. A page with climbing impressions and flat traffic may mean patients are receiving their answer directly on the results page and never clicking through. That pattern is not necessarily a failure, since being cited builds familiarity with a practice name, but it should not be read as traffic growth either.

Why This Matters for Medical Practices

Healthcare queries draw AI-generated answers at a high rate. Patients ask about symptoms, recovery timelines, procedure candidacy, and what a condition actually means, and those are precisely the questions generative features attempt to answer directly. Knowing which of your pages Google pulls from tells you where your practice has established credibility in the eyes of the model.

That knowledge has practical uses. If a single detailed post is carrying most of your generative AI impressions, it suggests the depth of that page is doing the work, and similar depth elsewhere may earn similar visibility. If your core service pages are absent from the report entirely, that is a signal about how thoroughly they answer real patient questions.

A Data Issue Worth Knowing About

Google confirmed a logging error that reduced reported impressions in the generative AI performance report in Search for data covering August 13 through 17, 2026. A related logging error affected the Discover performance report for August 13 data. Both were problems with how the data was recorded rather than real changes in how sites performed, so a dip across those dates does not reflect anything that happened in Search itself.

The practical step is to exclude that window when comparing periods, rather than treating it as the start of a decline. This is also a useful reminder that reporting problems and performance problems can look identical inside a dashboard. Confirming which one you are looking at should come before any decision about changing course.

How to Read the Report Without Overreacting

The most useful first step is establishing a baseline rather than drawing conclusions. Give the report several weeks of clean data, note which pages consistently appear, and resist redesigning a content strategy around a single week’s movement. Generative AI results also vary by model version, language, and account tier, so the underlying system is genuinely noisier than traditional rankings.

This report pairs well with the traffic questions practices were already asking. We covered the earlier difficulty of separating AI Mode traffic inside Search Console when no dedicated view existed. This report is the direct answer to that gap, though the missing click data means the picture is clearer rather than complete.

A New Way to Measure AI Search Visibility

The generative AI performance report is now available to every practice with a Search Console property, and it offers the first reliable look at how your pages perform inside AI-generated answers. Treat the current numbers as a starting baseline, set aside the mid-August logging window, and watch which pages hold their position over the coming months. If you would like help interpreting what the report shows for your practice, reach out to our team and we will walk through it with you.