Google's Generative AI performance reports in Search Console reached every website worldwide on 31 August 2026. For the first time, every site owner has a first-party number for how often their pages appear inside AI Overviews and AI Mode. What you still cannot see is how many people clicked through as a result.

That makes the report a visibility indicator, not a traffic metric. Used that way, it is genuinely useful. Used as a traffic report, it will mislead you.

📌 The short version:

  • The report counts impressions only: how often your URLs appeared in AI Overviews, AI Mode and generative AI features in Discover.
  • It breaks down by page, country, device and date. It does not show clicks or search queries.
  • AI Overviews and AI Mode are combined into one figure. Discover is reported separately.
  • The most valuable use right now: find pages that rank well in classic Search but earn zero AI impressions. Those pages are usually a structure problem, not an authority problem.

What Does the Generative AI Report Actually Show?

Google launched the reports to a limited set of sites on 3 June 2026, then extended them to all websites on 31 August. You will find them as an expandable section under the main Performance report in Search Console.

The report answers one question well: how often did a URL from your site appear inside a generative AI feature? You can slice that by:

  • Page, so you can see which URLs are being surfaced in AI answers and which are not.
  • Country, which matters for Australian businesses with any international traffic. Filter to Australia before you read anything else.
  • Device, useful because AI Overview layouts behave differently on mobile.
  • Date, with hourly, daily, weekly and monthly granularity.

The data is not new in the sense of being invisible before. AI feature impressions were already counted inside your overall Search performance totals, and they still are. What is new is that you can now isolate them.

What the Report Does Not Show

Four gaps matter, and each one changes how you should read the numbers.

  • No clicks. You can see that you appeared. You cannot see whether anyone came to your site from that appearance. Google has indicated click data may follow, but it is not there today.
  • No queries. You cannot see which searches triggered your AI appearances. You can see the page, not the question.
  • No split between AI Overviews and AI Mode. Both are rolled into a single figure, so you cannot tell which surface is driving the number.
  • Search Labs experiments are excluded. Anything Google is still testing will not appear.

📝 A correction to our earlier coverage: when we wrote about Search Console's AI performance reporting in July, early reporting suggested the new view would include AI-attributed clicks. The version that rolled out globally reports impressions only. We have since corrected that article. The click-side comparison has to come from your main Performance report and GA4 for now, not from the AI report itself.

Why an Impressions-Only Report Is Still Worth Having

A rising AI impression count looks like good news whether or not it is sending you customers, and that is the honest limitation here. But impressions do answer a question nothing else in Search Console could: is Google treating this page as a source worth citing?

That is a question about content authority and structure, and it is answerable today. It also gives you a baseline. When click data does arrive, you will want months of impression history to read it against, and the only way to have that history is to start recording it now.

Which Content Formats Are Pulling AI Impressions?

Across the published citation studies we have reviewed this year, three formats consistently attract the most AI Overview visibility:

  • Listicles and ranked lists. A May 2026 Evertune analysis of 25,000 AI-cited URLs found list-format pages were the single most-cited content type.
  • Comparison content. "X vs Y" and side-by-side evaluation pages match the way people ask AI systems to weigh options.
  • Definitional pages. Clear "what is" explanations are close behind, because they give an AI system a self-contained answer to lift.

There is a catch worth naming. The same formats that earn the most AI impressions are also the ones losing the most clicks, because the AI answer often satisfies the searcher. That is exactly why the impressions number on its own is not a success metric. We cover the click side of this in our zero-click search reality check.

The Most Useful Check: Strong Rankings, Zero AI Impressions

This is the analysis we are running for every SEO and GEO client this month, and it takes about half an hour.

  • Export your pages from the main Performance report for the last three months, filtered to Australia. Keep average position, impressions and clicks.
  • Export your pages from the Generative AI report for the same date range and country.
  • Join the two in a spreadsheet by URL.
  • Flag every page with an average position in the top five and meaningful classic impressions, but zero or near-zero AI impressions.

Those flagged pages are the interesting ones. Google already trusts them enough to rank them. It is not choosing to quote them. That gap is usually fixable.

Why Would a Page Rank Well but Not Get Cited?

Start with the innocent explanation. Some queries simply do not trigger AI features. Transactional and navigational searches, such as "plumber Parramatta" or a brand name, often show no AI Overview at all. If a page ranks for those, zero AI impressions says nothing about the content.

For informational and comparison queries, though, a ranking page with no AI visibility usually has a structural problem. The common ones:

  • The answer is buried. The page opens with background and takes five paragraphs to say anything citable. AI systems favour pages that answer first and elaborate second.
  • Headings describe topics rather than questions. "Our Approach" gives a retrieval system nothing to match. "How long does a bathroom renovation take in Sydney?" does.
  • Claims are vague. Specific figures, dates, named sources and clear definitions are easier to quote than general reassurance.
  • No supporting structure. Missing FAQ sections, missing schema and missing author signals all make a page harder to trust and parse. Our guide to schema markup for AI visibility covers the structured data side.

None of these require new content. They require reworking content that already ranks, which is the cheapest AI visibility win available to most sites. Our walkthrough on getting featured in Google AI Overviews goes into the rewrite pattern in more detail.

How to Report on It Until Click Data Arrives

We would add four lines to a monthly report, and resist the urge to add more:

  • AI impressions by page, Australia only, month on month.
  • The ranking-versus-citation gap list from the check above, and how many pages moved off it this month.
  • Classic clicks for the same pages from the main Performance report. If AI impressions climb while clicks on the same URLs fall, you now have evidence of substitution rather than a hunch.
  • Conversions from those landing pages in GA4. Fewer, better-qualified visits is a real possibility, and it only shows up if you are measuring outcomes. Our GA4 guide covers the setup.

The Bottom Line

The Generative AI report is the first official measure of whether Google treats your content as citable. It is not a traffic report, and it will not become one until click data arrives. Open it, filter to Australia, and compare it against your rankings. The pages that rank but are never cited are where your next round of content work should go. If you want help building that into a proper AI SEO and GEO program, that is what we do.

Sources: Google Search Central documentation, Search generative AI performance reports (updated 31 August 2026); Search Engine Roundtable, global rollout of Search Console generative AI reports and controls (1 September 2026); Search Engine Journal, Search Console AI reports worldwide (August 2026); Evertune AI citation analysis of 25,000 URLs (May 2026), as reported by AirOps. Figures verified against primary Google documentation where available.

Nauman Hashmi, Founder of DigiWolf
Written by
Nauman Hashmi
Founder & Principal Consultant, DigiWolf · Greater Sydney Area

Nauman has spent more than 15 years in digital marketing across Australia, the UAE and Pakistan, working in-house and agency-side before founding DigiWolf. He holds an MBA in Finance & Marketing from the Institute of Business Administration, and works directly with every DigiWolf client on AI search, SEO and paid media strategy.

Google Ads Search CertifiedGoogle Ads Video CertifiedGoogle Analytics IQAI Visibility Essentials, Semrush