Search Console Adds Generative AI Performance Reports

For the past two years the biggest blind spot in organic reporting has been AI. As AI Overviews and AI Mode took up more of the results page, site owners could watch impressions and clicks shift but had no clean way to attribute that movement to generative features. The new reports close part of that gap by carving out a dedicated view of AI visibility inside the tool most UK businesses already use to monitor search.

The underlying data is not entirely new. Google confirms that generative AI impressions have always fed into the overall Performance report, and they will continue to be tracked there. What changes is isolation: there is now a separate view showing only the AI surface, so you can see how much of your visibility comes from AI Overviews, AI Mode and the generative features in Discover rather than from the traditional ten blue links.

How the new reports work

The launch covers two dedicated reports, one for Search and one for Discover, sitting alongside the existing Performance reporting in Search Console. Google is rolling them out to a subset of websites first, so not every property will see them straight away. The phased release is deliberate: Google wants feedback from a smaller pool of site owners before making the reports widely available, and has built a dedicated feedback form into the interface.

Each report breaks visibility down across familiar dimensions. Impressions show how often URLs from your site appeared inside generative AI features on Search and Discover. The Pages view lists which specific URLs surfaced, which is the detail most teams will reach for first. Countries data splits visibility by market, and a Devices breakdown (available for Search results) shows which devices people were using when your pages appeared. Dates round it out with hourly, daily, weekly and monthly granularity, so movements can be tied back to algorithm shifts, content changes or seasonal demand.

One nuance matters at launch: the reports centre on impressions rather than clicks. That reflects how AI features behave. A page can be cited or summarised inside an AI Overview without the user ever clicking through, so impression share becomes the leading indicator of whether your content is being drawn into generative answers. Google has said it plans to add further metrics over time, informed by the feedback it gathers during the rollout. For background on how AI features are presented in results, Google maintains its own documentation on AI features in Search.

What the reports let you see

Consider a UK professional services firm that has watched organic clicks soften over the past year while impressions held steady. Until now, the most likely explanation, that AI Overviews were answering more queries directly, was a hypothesis with no supporting data inside Search Console. With the generative AI report, the firm can confirm whether its pages are appearing in AI features, which URLs are being pulled in, and whether that exposure is concentrated in particular markets.

A retailer publishing content across both Search and Discover gains a second angle. Discover has long been an opaque channel, and adding a generative AI view for it means publishers and ecommerce brands can finally separate standard Discover impressions from AI-driven ones. For a brand investing in editorial content to support product discovery, that distinction shows whether the AI surface is rewarding the work.

The Pages dimension is where most practical decisions will start. By identifying the exact URLs that appear in AI features, teams can study what those pages have in common: clear structure, a direct answer near the top, genuine topical depth, supporting schema. Those patterns then become a template for the rest of the site, turning a reporting feature into a content brief.

What this means for UK site owners

For UK agencies and in-house teams, the headline is that AI visibility is becoming measurable inside a free, first-party tool rather than something inferred from third-party trackers or estimated from traffic dips. That matters, because the measurement gap has been the single hardest part of briefing clients on AI search. We covered the scale of that problem in our analysis of tracking AI search traffic in GA4, where the core issue is that most AI-referred sessions land as direct traffic with no usable referrer. Search Console approaches the question from the visibility side rather than the traffic side, which makes the two data sources complementary rather than redundant.

The practical move is to fold the new report into existing routines rather than treat it as a novelty. If you already pull a monthly Search Console export for clients, add the generative AI view and track impression share on the AI surface over time. That gives a defensible baseline before the next core update reshapes how much of the results page AI occupies. For teams building a longer-term plan, it slots neatly into the framework we set out in our UK playbook for AI search optimisation.

None of this replaces the fundamentals. Appearing in AI features still rests on the same foundations as ranking well in traditional results: authoritative content, clean technical SEO, and pages that answer the question directly. The report tells you whether that work is paying off on the AI surface. It does not change what the work is.

Practical considerations

A few caveats sit against the optimism. The first is access. Because Google is releasing the reports to a subset of properties, many UK site owners will not see them yet, and there is no published timeline for full availability. If the report is missing from your Search Console, that is expected during the testing phase rather than a sign of a configuration problem.

The second is interpretation. Impressions inside AI features are not the same as clicks, and they are certainly not the same as conversions. A high AI impression share confirms that your content is being surfaced, but the commercial value of that exposure depends on whether users still visit your site, and that behaviour is exactly what the wider industry is still working to quantify. Treat the new numbers as a visibility signal, not a revenue forecast.

The third is volatility. AI features on Search are changing quickly, and the data behind these reports will move as Google adjusts how AI Overviews and AI Mode are triggered. Early figures should be read as directional, and comparisons across short windows during the rollout are likely to be noisy. Building a reliable trend line will take months. The general mechanics of the Performance report, including how impressions are counted, are set out in Google's official Search Console help documentation.

How to fold it into reporting

A short routine makes the report pay back faster. First, check availability across your portfolio and note which properties have the reports live, so you know where AI data can already inform decisions. Second, record a baseline this month: total AI impressions, the top ten pages appearing in AI features, and the split by country. Third, cross-reference those pages against your traditional Performance report to see where AI exposure and classic ranking overlap and where they diverge. Pages that perform on both surfaces are your strongest assets. Pages that appear in AI features but rank poorly in standard results point to a content or technical gap worth closing.

FAQ