What are GSC AI performance reports?
On 3 June 2026 Google announced dedicated performance reports for generative AI features in Search and Discover via the Search Central blog. Concretely: separate views of impressions in AI Overviews, AI Mode and Discover AI alongside the existing standard performance report.
Data starts from 18 May 2026. That means even by the end of June you have at most about six weeks of history to work with. Enough for a first baseline, too short for seasonal patterns.
An important difference from the standard Performance report: these new reports only measure appearances within AI features, not within the traditional ten-blue-links results. A query where your site appears both in the blue links and in an AI Overview gets two separate registrations — one in each report.
What the report does and does not show
What is included: impressions within AI features, page name, country, device type and a time series. Filtering works as in the existing report.
What is missing: clicks and CTR data. Google has indicated clicks data will be added later, with no date. Query-level breakdown is also missing — you cannot see which user question triggered your AI Overview impression.
For SME reporting this means: you can prove visibility within AI features but you cannot yet directly attribute traffic from them. Combine with the GA4 ai-assistant medium Google added in May — that measures traffic from AI tools to your site, even though GSC does not yet measure click-through from AI Overviews themselves.
Getting access to the limited beta
The rollout is a phased beta. Not every Search Console property has the reports available yet. No sign-up form, no waitlist — Google rolls out gradually.
Check your access: log into Search Console, pick a property, and look under Performance for a section called 'Search Generative AI' or similar. If absent, your property is not yet in the beta.
Fallback for non-beta properties: use the standard Performance report plus the GA4 ai-assistant medium as a proxy for AI visibility. Expect general availability within the next three to six months based on Google's rollout history with similar features.
Interpreting the data: four scenarios for SME sites
Scenario A — high AI Overview impressions, flat organic clicks: your content is being cited in AI answers but users are not clicking through. Response: shorter, stronger opening paragraphs that build curiosity rather than give the answer away.
Scenario B — low AI impressions, strong organic position: your page ranks in blue links but is not chosen by AI for citations. Response: review extractability — clear question structure, schema markup, concise definitions.
Scenario C — AI Mode impressions rising month-over-month in double digits: positive signal. Your entity clarity and topical authority are in order. Response: extend the same approach to sister content.
Scenario D — zero AI impressions over a longer period while organic rankings hold: likely a block. Check robots.txt for AI crawler blocks (Google-Extended, ChatGPT-User), check llms.txt if present, check meta noai tags.
Five tactical actions for SME agencies
1. Measure the baseline as soon as data arrives. Export impressions per page from the first data point (18 May). This becomes your comparison base for future optimisations.
2. Identify which pages already trigger AI Overviews. Often these are FAQ-rich pages, definition content or step-by-step guides. Lift that pattern to other pages.
3. Run a robots.txt and llms.txt audit. Many sites accidentally block AI crawlers through legacy rules or excessive security caution. An accidental Disallow for Google-Extended can keep your AI visibility at zero.
4. Implement or upgrade schema markup for extractability. FAQPage, HowTo and BreadcrumbList in particular are AI-citation-friendly. Combine with answer-first structure: every H2 opens with a concise definitive statement.
5. Build a client reporting template that links GSC AI data to GA4 ai-assistant traffic and organic traffic. Three-layer story: visibility within AI, traffic from AI tools, and traffic from traditional search results. For SME clients this is a fresh selling point.
What is coming in the next six months
Google has stated explicitly that more metrics will be added. Expect clicks data within three to six months, possibly followed by query-level breakdown later in 2026.
General availability for all properties will likely roll out in stages — large properties first, then SME-tier properties. Build reporting templates that already contain these data fields, even if they stay empty until access is granted.
The bigger shift: on 15 May 2026 Google retired GEO and AEO as separate optimisation disciplines with the message 'optimizing for AI Search is SEO done well'. The new reporting system reinforces that — there is no separate AI optimisation measurement anymore, there is only Search Console with an extra view for AI features.
Want to know if your site is already visible in AI Overviews?
Request an AI visibility audit →