Google Search Console now gives site owners a dedicated view of visibility inside Google’s generative AI search features. The Generative AI performance report covers AI Overviews and AI Mode, showing when links from a site appear within those experiences.
Google has solved part of the visibility problem before solving attribution. The report provides first-party AI impression data, but it still does not show AI-specific clicks, queries, CTR, or a separate split between AI Mode and AI Overviews. For SEO teams, the report is useful mainly because it shows where AI exposure exists and how it changes over time.
What the Google Search Console AI Report Actually Measures
Google announced the Search Generative AI performance report on June 3, 2026. According to Google’s Search Console announcement, the Search version includes AI Overviews and AI Mode, while Search Labs experiments are excluded.
The core metric is impressions. Google records an impression when a link to a page from your site appears to a user inside a supported generative AI Search feature. Those impressions can then be broken down by page, country, device, and date.
An impression records exposure, not a visit, so the report works better as an AI visibility dataset than as a traffic report. A page can appear frequently inside generative AI results while producing very different click behavior from a conventional organic listing.
What You Can See in the Generative AI Performance Report
Google currently lets site owners analyze four dimensions:
- Pages — which canonical URLs receive generative AI impressions;
- Countries — where that visibility comes from;
- Devices — differences across desktop, mobile, and tablet;
- Dates — how AI visibility changes over time.
The Pages dimension is probably the most useful for content analysis. If a small set of URLs repeatedly appears in AI Search, those pages can be compared for subject matter, freshness, sourcing, format, and search intent. A broader pattern can matter even more than one large number: if ten pages in the same topic cluster begin receiving AI impressions, Google may be expanding its use of that section of the site.
Country and device data help prevent property-wide totals from becoming misleading. A business targeting the United States may find that much of its AI visibility comes from other markets, while mobile and desktop impressions can follow different patterns. The report does not explain those differences, but it gives enough information to separate them before drawing conclusions.
Dates are most useful over weekly or multi-week windows. Small sites can produce dramatic percentage changes from a handful of impressions, so daily movement is easy to overinterpret. Longer comparisons are more useful after a major content update, new publication, indexing event, or broader Search change.
Access is also still being rolled out. The official Generative AI performance report documentation says the report is available to a subset of website owners, so its absence does not automatically mean a site has zero AI visibility.
What the Search Console AI Report Still Does Not Show
The missing metrics shape how far the report can be taken. A normal Search Console workflow revolves around impressions, clicks, CTR, queries, and position, while the dedicated AI report exposes only part of that set.
There are no AI-specific clicks
Google can count external clicks from AI Mode and AI Overviews under its normal Search measurement rules, but the dedicated Generative AI report does not isolate them.
If a page receives 5,000 generative AI impressions, Search Console cannot tell you how many visits came from those particular appearances. The clicks may still be present inside standard Web Search reporting; they simply cannot be tied back to the AI-only view.
You cannot calculate a reliable AI CTR
Without AI-specific clicks, there is no clean way to calculate AI CTR from the dedicated report. Dividing total Web Search clicks by generative AI impressions would mix a broad click total with a narrower impression set.
That kind of calculation can produce a convincing percentage that does not actually describe AI search behavior. A page with fewer AI impressions could generate more external traffic than another page with much greater visibility, and the current report cannot distinguish between those outcomes.
Search Console does not reveal AI queries
The missing Queries dimension is probably the biggest limitation for SEO work. Search Console can show that a page appeared hundreds of times inside generative AI Search, but it does not show what users searched or asked when those impressions occurred.
That removes the usual connection between page and intent. A page about AI Mode might receive 800 generative AI impressions, for example, but the report cannot tell you whether those impressions came from searches about AI Mode usage, ecommerce research, citations, shopping behavior, or another topic entirely.
Traditional query data often tells you whether Google understands a page narrowly or is testing it across a wider cluster. The AI report hides that layer, so page-level visibility becomes useful without becoming fully interpretable.
Dedicated AI position data is missing too
The report does not provide a separate average position metric for generative AI visibility. In practice, this omission is less restrictive than the missing query data because conversational AI results do not fit cleanly into the traditional ten-blue-links model.
Google still applies position rules to AI features within standard Search reporting, but there is no dedicated AI position view. For most site owners, page coverage and impression trends will probably be more useful than trying to recreate classic rank tracking around AI Mode.
AI Mode and AI Overviews Are Combined
Search Console combines AI Mode and AI Overviews in the same Generative AI performance report. There is no dimension showing how many impressions came from each product separately.
A total of 1,000 generative AI impressions therefore should not be described as 1,000 AI Mode impressions. The total may come from AI Overviews, AI Mode, or both.
That distinction is not trivial. Current Google AI Mode statistics show that AI Mode involves longer prompts, more follow-up activity, multimodal input, planning, and comparison behavior. Search Console does not expose that product separately, so a visibility increase cannot automatically be attributed to AI Mode itself.
For publishers, the safest interpretation remains page-level: a URL appeared more often inside Google’s generative AI Search environment. Product-specific conclusions require data the report does not currently provide.
Do Not Add AI Impressions to Web Search Impressions
Generative AI impressions are already included in the standard Web Search data shown in Search Console. The dedicated report is another view of part of that existing Search visibility, not an extra pool of impressions.
If Web Search reports 10,000 impressions and the Generative AI report shows 1,500, the total is not 11,500. Those 1,500 AI impressions are already part of the broader Web Search figure.
For dashboards and client reports, double counting is the immediate risk. The cleaner approach is to treat generative AI impressions as a subset of organic Search exposure and report them separately without adding them back into the total.
What an AI Impression Actually Means
An AI impression means that a link from the site appeared to a user in a supported generative AI Search feature. It does not show whether the user clicked, remembered the brand, trusted the source, became a lead, or made a purchase.
That distinction matters more in AI Search because users can continue researching inside Google. They may refine the question, compare options, or gather enough information from the generated answer that an external visit becomes unnecessary.
For reporting, “AI visibility” is therefore the more accurate label unless analytics confirms visits. Search Console measures exposure first; business value has to be established elsewhere.
How to Use Google Search Console AI Visibility Data
The report becomes most useful when several page-level patterns are compared instead of focusing on one property-wide total. A publisher can see which content types begin receiving AI impressions, whether new pages appear shortly after indexing, and whether visibility spreads across a topic cluster.
It can also highlight differences between sections of the site. Detailed research pages may gain AI exposure while broad commercial pages remain absent, or one category may begin appearing consistently while another does not. Those differences can guide further content analysis even though the report cannot explain why they occurred.
Multi-week comparisons are usually safer than daily ones. If a page is updated with better sourcing, fresher data, or a clearer structure and AI impressions subsequently rise, that is worth noting, but it is not proof that one edit caused the change. Google’s ranking systems and AI surfaces continue changing at the same time.
A Better Way to Read AI Search Performance
The easiest reporting mistake is asking one tool to answer a business question it cannot answer. The main datasets involved in AI Search measure different stages of the same journey.
| Data source | Best use |
|---|---|
| GSC Generative AI report | AI Search visibility |
| Standard Search Performance report | Overall Google Search impressions, clicks, queries, CTR and position |
| Web analytics | Visits, landing-page behavior and conversions |
| CRM or sales data | Lead quality, pipeline and revenue |
A rise in Generative AI impressions tells you that visibility increased. It does not tell you whether traffic rose, whether visitors converted, or whether the resulting leads were commercially useful.
Keeping those layers separate produces a more defensible analysis than combining them into a homemade “AI performance score.” The reporting becomes slightly less tidy, but much more accurate.
The August 2026 Reporting Bug Matters for Comparisons
August data needs a specific warning. Google’s official Search Console data anomalies record says a logging error reduced reported impressions in the Generative AI Search performance report from August 13 through August 17, 2026.
Google says the problem affected logging rather than actual Search visibility. A drop during those dates should therefore not be treated automatically as evidence that a site lost AI exposure.
This can distort month-over-month comparisons and short before-and-after analyses. If August 13–17 falls inside the baseline, a later increase may look larger than the underlying change really was, so those dates should be annotated or excluded when evaluating AI impression trends.
What Search Console Still Cannot Tell You About AI Traffic
The dedicated report gives site owners something they did not have before: first-party visibility data from AI Overviews and AI Mode, broken down by page, country, device, and date. That is already enough to identify where AI exposure is appearing and whether it is expanding.
The unanswered questions are mostly about attribution. Search Console does not yet expose AI-specific clicks, CTR, queries, or an AI Mode versus AI Overviews split, so the report cannot independently explain traffic or commercial value.
Google said when announcing the report that it is considering additional metrics over time. Until that happens, the most useful approach is to treat the Generative AI report as one layer of measurement, combine it with standard Search Console, analytics, and business data, and avoid turning AI impressions into claims the underlying data cannot support.