On June 3, 2026, Google launched the Generative AI features report in Search Console — the first official view of how often your pages appear in AI Overviews (the AI summaries in Google Search results), AI Mode (the conversational search mode), and generative features in Discover. The beta version comes with serious limits, though: it shows impressions only, no clicks, no CTR, no queries. This guide covers where to find the report, how to segment it, and — most importantly — how to read those impressions correctly: as a signal of presence in AI features, not as traffic.
What Google launched
The Generative AI features report has been available in Search Console since June 3, 2026 (Beta), and as of publication it is still rolling out to a subset of sites (Google says access is being granted gradually). Not every Search Console property has it yet. It has two separate views:
- Search — AI Overviews and AI Mode in Search.
- Discover — generative AI features in Discover.
What the report shows
Metrics and dimensions in the Generative AI features report
- Impressions The headline metric — how many times your URLs appeared in generative AI features in Search and Discover.
- Pages Which specific URLs appeared in AI features.
- Countries Visibility broken out by country.
- Devices Mobile versus desktop — Search view only.
- Dates Hourly, daily, weekly, and monthly intervals for tracking the trend.
The critical limitation: no clicks
The most important thing to know about this report: version 1 tracks impressions only. No clicks, no click-through rate (CTR), no position, no queries.
Google doesn’t publish clicks, CTR, position, or queries in this version of the report. In practice, that means you can’t read traffic or post-click performance out of it. One likely explanation is that people interact with AI answers differently — those answers often resolve the question right in the interface, so a click isn’t necessarily the main outcome of the interaction. We cover that effect in detail in the article on the zero-click effect.
How to read impressions correctly
The right way to read an impressions-only report is to treat impressions as a signal of presence — a signal that your URL is showing up in Google’s AI features — not as a traffic metric. They can be a useful signal that your content is relevant inside AI features, but they aren’t direct proof of quality, traffic, or business value.
-
High AI impressions = your URL shows up more often in Google AI
A page with high AI impressions appears frequently in Google's AI features. That's useful information in itself about which content is visible inside AI features — even if it isn't yet proof of quality or value.
-
Compare AI impressions against organic clicks
Take the same URL and compare its AI impressions in the new report with its organic clicks in the classic Performance report. The gap between them tells you whether the content also brings traffic.
-
Separate what's visible from what brings traffic
High AI impressions plus low organic clicks means content that's visible in AI but isn't sending traffic (the zero-click effect). High AI impressions plus high clicks means content that works in both layers.
-
Track the change over time
The report supports hourly through monthly intervals. Watch whether AI impressions are rising or falling — and tie that back to your content changes or to changes on Google's side.
How to find and segment the report
-
Confirm you have access to the report
The report has been rolling out since June 2026. Under Performance in Search Console, check whether a Generative AI features section is available. If it isn't there yet, stick with a manual test for now.
-
Go to Performance and switch to Search
If the report is available in your interface, open Performance in Search Console and then Generative AI features. Switch between the Search view (AI Overviews + AI Mode) and Discover. For AI search, focus mainly on Search.
-
Filter by page, country, and time
Filter by specific URLs, by country, and by date. The default Performance report mixes all search types together, which masks AI trends — so use a separate filter or export for AI features only.
-
Track the AI report on a regular schedule
Build your own recurring view on top of the AI report — a data export, for example — segmented by page, country, and date. That's the foundation for tracking how your site's AI visibility develops.
What the GSC AI report can’t do
The report is a strong new layer, but you need to know where it stops.
| Aspekt | Shows | Doesn't show |
|---|---|---|
| Impressions | Yes — how many times your URLs appeared in Google AI | — |
| Clicks and CTR | — | No — version 1 doesn't track clicks |
| Queries and position | — | No — no query-level data |
| Revenue and conversions | — | No — it won't tell you whether an impression led to a sale |
| AI outside Google | — | No — Google AI only, not ChatGPT, Perplexity, or Claude |
Where the GSC report fits in AI measurement
Search Console is one layer. For the full picture of your AI visibility and what it’s worth, you need more than one source.
| Aspekt | What you measure | What you measure it with |
|---|---|---|
| Google AI impressions | How often you show up in Google AI | Search Console — Generative AI features report |
| Traffic from AI | How many people came to your site from AI | GA4 (AI referrals, partially) + UTM parameters |
| Citations outside Google | How ChatGPT, Perplexity, and Claude cite you | A manual test across platforms or dedicated tools |
| Quality of AI traffic | Whether an AI visitor converts | GA4 + UTM + conversion tracking |
The most common mistakes
Reading impressions as traffic
High AI impressions don’t mean high traffic. The report doesn’t show clicks. Treat impressions as traffic and you’ll overestimate what AI visibility is actually worth.
Relying on GSC alone
The report covers Google AI only. ChatGPT, Perplexity, and Claude aren’t in it. Measure only in Search Console and you’re seeing one slice of the AI ecosystem.
Leaving the data in the default report
The default Performance report mixes all search types together, and AI trends disappear into the aggregate. Without a separate filter on AI features, the specific picture is lost.
Not combining it with organic clicks
AI impressions on their own tell half the story. Only comparing them with organic clicks for the same URL shows you whether the content brings traffic or is merely visible.
Key takeaways
- Google launched the Generative AI features report (Beta) in Search Console on June 3, 2026 — the first official data on Google AI visibility.
- It shows impressions, pages, countries, devices, and dates — across two views (Search and Discover).
- It doesn’t show clicks, CTR, position, or queries — version 1 tracks impressions only.
- Read impressions as a signal of presence in Google’s AI features, not as traffic; compare them with organic clicks.
- The report covers Google AI only — for ChatGPT, Perplexity, and Claude you need a manual test.
- Supplement GSC with GA4, UTM, and a manual test — it’s one layer, not the full picture.
Want to set up complete AI visibility measurement across Search Console, GA4, and manual testing? This site is run by Sniper Design — our AI visibility audits connect Search Console data, GA4, and cross-platform manual testing, then show you where a change will have the biggest impact.
For transparency: we run our AI audits with measurement across all these layers (Search Console, GA4, manual testing). The description of the report is based on public Google Search Central documentation as of June 2026; the features are in Beta, and both their scope and availability can change. We describe how the content on this site is produced on the author page.
