Search Console Now Reports AI Overview and AI Mode Performance

For the last two years the hardest question in AI search has been a measurement one. You could do the work to be cited in AI Overviews and AI Mode, but you could not see whether it paid off. The visibility was real and the data was missing. On June 3, 2026, Google started to close that gap by adding dedicated generative AI performance reports to Search Console. This is a genuine step forward, and like most first steps it comes with limits worth understanding before you read too much into the numbers.

This piece looks at what the reports actually show, where the early rollout stops, how to read the data calmly, and what it does and does not change about the work of getting found by AI.

What Google Announced

The new reports give you a dedicated view of how your site appears inside the generative AI features on Search, specifically AI Overviews and AI Mode, plus the generative AI features in Discover. Until now that activity was folded into your ordinary Search performance, with no way to separate it out. Now it has its own reporting surface.

The data covers five dimensions: impressions, pages, countries, devices for Search, and dates with granularity from hourly through monthly. There are separate views for Search and for Discover. In plain terms, you can finally see which of your pages are showing up in AI generated answers, in which countries, and how that trends over time.

Alongside the reports, Google is rolling out a control that lets site owners block their content from appearing in or around these AI features, as links or for grounding. That opt out is a separate decision from measurement, and for most sites it is not one you will want to reach for. We will come back to it below.

The Limits of the Early Rollout

This is the part to be clear eyed about, because the headlines make the feature sound more finished than it is today. The reports are rolling out to a subset of sites, and at launch that subset is a small group of UK site owners. If you do not see the report in your Search Console yet, that is expected, not a setup problem on your end. Google has said it will expand from here.

The data also starts on May 18, 2026 and moves forward from there. There is no deep history to look back on, so early readings are a short window rather than a trend you can lean on. And the dimension on offer is impressions, not the full click and position picture you may be used to from classic Search. Knowing you appeared in an AI answer is useful, but it is not the same as knowing what that appearance earned you.

None of this makes the feature less welcome. It makes it early. The right posture is to treat the first numbers as a signal that the measurement gap is starting to close, not as a finished dashboard to optimize against.

How to Read the Data Without Overreading It

A new report is exciting, and excitement is exactly when misreadings happen. A few steady habits keep the early data honest.

Read impressions as presence, not as proof of traffic. An impression here means your page was part of an AI generated answer. That is worth knowing, but the relationship between an AI impression and a real visit is still loose, and Google is not yet handing you a clean click and attribution model for it. Resist the urge to translate impressions directly into value.

Watch which pages appear, more than how many. The most useful thing in the early data is the list of pages that surface in AI answers at all. That tells you what AI systems currently consider answer worthy on your site, which is a far more actionable signal than a raw count from a two week window.

Give it time before drawing conclusions. With data starting in May 2026 and a partial rollout, any trend line is short. Note what you see, but do not restructure your content strategy around a fortnight of impressions. The honest move is to let the window grow first.

The Blocking Toggle, and When It Makes Sense

The opt out control deserves a calm word, because it is easy to reach for out of unease about AI using your content. For the large majority of sites, leaving it off is the right call. Appearing in AI Overviews and AI Mode is visibility, and blocking yourself removes you from answers your competitors will still be in. The measurement reports exist precisely so you can understand that visibility rather than fear it.

There are narrow cases where opting out is reasonable, for example content you are contractually unable to expose for grounding, or sections of a site where AI summarization genuinely conflicts with how the content must be used. Those are specific decisions, not a default. If you are reaching for the toggle simply because AI search is unfamiliar, the better response is to measure first and decide from evidence.

What This Changes for GEO, and What It Does Not

It is tempting to treat a new report as a new strategy. It is not. What changed is that a slice of AI visibility became measurable. What did not change is the work that earns that visibility in the first place.

The signals that get you into AI answers are the same ones we covered in the GEO playbook: clean, crawlable, well structured content, JSON-LD that describes your entities, an llms.txt that AI assistants actually read, and robots.txt rules that do not accidentally lock AI crawlers out. The new report does not replace any of that. It gives you a way to watch the result of doing it well. Think of it as a thermometer, not a treatment.

If anything, better measurement raises the value of the groundwork. When you can finally see which pages surface in AI answers, the next question is always why those and not others, and the answer almost always lives in structure, clarity, and technical health. The report points at the work. It does not do it. The wider context for why Google is building this at all is in our look at Google’s own AI optimization guidance.

Where Seodisias Stands on This

To be straight about it, this is a Google reporting feature, and Seodisias does not pull data from your Search Console. The two sit on different sides of the same problem. The Search Console report tells you whether you appeared in an AI answer. Seodisias tells you whether your pages are built so an AI engine can read, parse, and trust them in the first place.

That is what the AI Ready analysis is for: it checks structured data, content structure, llms.txt, and the technical signals that decide whether your content is even legible to an AI system, as part of a normal crawl. The new report and a crawl based readiness check are complementary, not competing. One measures the outcome, the other inspects the cause. Used together, you can see where you show up and understand why, which is more than either gives you alone.

Conclusion

Search Console finally reporting on AI Overviews and AI Mode is good news. For two years the main complaint about AI search was that you were working in the dark, and this lights a candle. Just keep the candle in proportion. The rollout is partial, the history is short, the metric is impressions rather than clicks, and the data is early. Read it as presence, watch which pages surface, and let the window grow before you conclude anything.

Underneath the new dashboard, the job is unchanged. Clean, structured, crawlable content is still what gets you into AI answers, and a crawl that checks whether your pages are AI ready is still the most direct way to act on what the report shows you. Measure the outcome in Search Console, fix the cause in your own pages, and the two will start to move together.