Google’s New Generative AI Performance Report in Search Console: What to Track First

Google’s new Search Console generative AI performance report just changed how we measure visibility in AI search. If you’ve struggled with sparse keyword data, this generative AI performance report gives you a reliable early signal to guide content and SEO investments.

What changed in June 2026, and why does the Search Console generative AI performance report matter now?

On June 3, 2026, Google launched dedicated Search Generative AI performance reports for Search and Discover that isolate your site’s visibility inside AI features like AI Overviews and AI Mode. The new report focuses on impressions with dimensions for pages, countries, devices, and dates, and Google confirmed a gradual rollout.

The support documentation clarifies that the generative AI performance report includes impressions for AI Overviews and AI Mode, draws its data from the Web search type in the main Performance report, and allows you to export both chart and table data. It also notes that experiments in Search Labs are excluded from these reports.

Industry coverage underscores two important caveats: the rollout is incremental and the new report does not include clicks yet. That means impressions are your primary leading indicator in AI features.

an SEO analyst reviewing the Google Search Console dashboard on a laptop at a modern office desk, with charts and country filters visible on screen

What exactly counts as an AI impression in Google Search?

In the report, an AI impression is counted when a link to your site is shown to a user within a generative AI feature on Google Search. This aligns with how Search Console defines impressions and how the performance reports aggregate visibility.

Google’s John Mueller recently explained that impressions are tied to actual links: if a user must expand an AI panel before your link appears, it only counts once the link is visible. He also reiterated that click data is not part of the current AI reporting, and noted that links within an AI Overview share a single position. This helps explain why observed appearances may exceed counted impressions.

The report aggregates totals by property in the chart and by your chosen dimension (pages, countries, devices, dates) in the table, with familiar limitations like 1,000-row exports—so workflows you already use for performance reports in Search Console will feel familiar.

Why is impression data a powerful leading indicator for teams with sparse keyword data?

When you lack statistically significant query samples, impressions become your earliest read on whether the site appeared in generative AI. Because the new report isolates AI Overviews and AI Mode, you can separate AI-driven visibility from classic blue links and build hypotheses faster—well before click patterns stabilize.

Practically, impression deltas surface which URLs and content types are being cited or linked inside AI features. That’s critical for organizations where the long tail dominates, branded demand is volatile, or seasonality masks signal in traditional query-level data. Treat AI impressions like “awareness exposure” within generative results, not as intent-validated demand.

What should you track first in the generative AI performance report?

Pages with rising AI impressions: Flag URLs gaining visibility in AI Overviews or AI Mode and compare on-page answers, structure, and entity coverage to peers.
Country/device splits: Watch whether mobile or specific markets lead AI visibility shifts; that can inform language, snippet length, and answer formatting strategy.
Date trends vs. content changes: Annotate major updates to map cause and effect as impressions move. Industry reporting indicates no click metrics yet, so impressions are your main time-series signal.

How do you read AI Overviews vs. AI Mode in this report?

The generative AI performance report breaks out impressions for AI Overviews and AI Mode so you can evaluate both surfaces independently. Since links inside an AI Overview share a single position and may require expansion to register an impression, variance between observed and counted visibility is expected. Align your analysis to link exposure, not mere mention.

Because the report’s data lives within the Web search type, your AI visibility is trackable alongside overall Search trends—helpful when comparing “new search” surfaces to traditional SERPs. Use dimension filters to isolate each AI feature before you compare to the all-up Performance report.

What are the current limitations of AI performance reports—and how should you mitigate them?

First, Google states this is a gradual rollout and some properties won’t see the report until access expands. Second, click data isn’t available yet, which means you should avoid CTR or revenue modeling for AI features at this time. Third, Search Labs experiments don’t populate these reports, so don’t expect all experimental surfaces to appear.

Fourth, Discover AI data has had known logging anomalies this summer; if you’re measuring Discover’s generative AI features, annotate June 24, 2026 (and other listed dates) to prevent misattribution. Finally, the usual Search Console limits (row caps, preliminary data) apply—plan exports and QA accordingly.

How can you turn AI impressions into a predictive signal when keyword data is thin?

Use a repeatable, small-sample-friendly framework that treats impressions as directional evidence rather than precise outcomes. Stabilize signal with segmentation and smoothing, then layer hypotheses you can test quickly in content.

Segment first: Group pages by template, intent, or entity cluster to reduce noise and spot where the site appeared in generative AI most consistently.
Use moving medians: Smooth 7–14 day medians to cope with preliminary data and irregular triggering of AI Overviews.
Compare adjacent cohorts: For each update, compare optimized pages to similar unoptimized pages over the next two weeks.
Triangulate with position: While AI position is shared inside Overviews, your classic rankings often move in parallel; divergences suggest cannibalization or summarization without linkage.
Adopt control charts: Control limits help you separate real change from natural variance when sample sizes are small.

7 Steps to Make AI Impressions Your Leading Indicator

Where do you find the new report, and how does it relate to other performance reports in Search Console?

You’ll see a dedicated Generative AI performance report entry in Google Search Console, with familiar tabs for Pages, Countries, Devices, and Dates. The data is included in the overall Web search type within the Performance report, so you can use side-by-side filters to reconcile AI visibility with classic traffic.

Because this lives within the existing reporting stack, reports in Search Console can be exported like your other performance reports. Those exports are essential for smoothing sparse data and building time-series controls across multiple cohorts.

Should you use the Search generative AI control to opt out—or stay in?

Google added a Search generative AI control in settings to include or exclude your site’s links and content from AI Overviews, AI Mode, and generative AI features in Discover. Excluding removes your eligibility for impressions and traffic from these features but does not act as a ranking signal for other parts of Search.

For most organizations, staying opted in is prudent while measuring risk and reward with the generative AI performance report. If sensitive sections must be excluded, use the property- or child-property level controls and monitor impression impact after changes propagate.

Which quality and policy guardrails should guide your AI visibility strategy?

Google’s Search spam policies apply to generative AI features on Google Search, including AI Overviews and AI Mode. Avoid manipulative tactics intended to force citations or links in AI answers; instead, strengthen content clarity, entity coverage, and citation-worthy sources. Think “answer equity” over “prompt gaming.”

Competitively, Microsoft’s Bing Webmaster Tools has previewed richer AI reporting fields like Intents, Topics, and Citation Share. That industry direction suggests AI performance reports will keep maturing, so establish your baseline now to benefit from future metrics.

a product manager and content strategist standing at a whiteboard mapping user questions and entity relationships with sticky notes

How do AI impressions help you prioritize action when keyword data is scarce?

Use AI impressions to find content that’s nearly “citation-ready” but under-linked in AI panels. Elevate concise answers, add authoritative references, and clarify entities so your URL is more likely to be exposed when an Overview expands. Then observe whether impressions lift per page cohort after the update.

Where you see impressions building without parallel click growth in classic search results, pursue internal linking and snippet optimization. Remember: current AI reports emphasize exposure, not engagement, so think top-of-funnel visibility now and revenue attribution later when Google surfaces more metrics.

What’s a 30‑day starter plan to operationalize the Search generative AI performance report?

Week 1 — Baseline: Export 90 days from the generative AI performance report by Pages, Countries, and Devices; build cohorts by template and intent; create control charts with 7–14 day medians.
Week 2 — Hypotheses: For top-cohort URLs, standardize intros, FAQs, and summary boxes; ensure authoritative references and schema where appropriate; document changes in annotations.
Week 3 — Compare: Track impression shifts for optimized vs. control pages; investigate AI Overviews vs. AI Mode gaps; refine entity coverage and on-page answers.
Week 4 — Scale: Roll successful patterns to adjacent cohorts and set a monthly review cadence, keeping an eye on data anomalies that could impact Discover AI reporting.

Key takeaways for website owners using Google Search Console today

Impressions are your leading indicator for AI Overviews and AI Mode until clicks arrive in reporting.
Segment and smooth to make sparse keyword data actionable across pages, markets, and devices.
Track exclusions carefully using the Search generative AI control, and measure effects on AI visibility post-change.
Annotate anomalies—especially the June 24, 2026 Discover logging issue—so you don’t misread trend lines.
Follow Search Central guidance and reputable industry reporting as Google iterates on AI features and metrics.

Where this is headed next—and how to stay ready

Google has signaled that it’s collecting feedback and may add additional metrics over time, so capturing baselines now will pay dividends when richer data—like clicks or more granular positions—arrives. Until then, use impressions as a directional compass and treat improvements as indicators of stronger answer equity in the new search experience.

The faster you can turn those early signals into content and UX refinements, the more resilient your visibility will be across generative AI features in Discover, AI Overviews, and AI Mode. That’s the practical path to sustainable performance as generative AI search evolves.

Need help building an impression-led AI search measurement plan?

If your team is dealing with sparse keyword data and you need a faster, more trustworthy signal from the Google Search Console AI performance reports, Xadira can help. We’ll set up exports, smoothing, and cohort comparisons; align visibility lifts to on-page improvements; and design a cadence that scales. Talk with our specialists at https://xadira.com to turn AI impressions into a winning roadmap.