For two years, marketing teams watched their organic CTR compress while Google's AI Overviews expanded across the SERP. The data gap was maddening: you could see impressions climbing and clicks falling in Search Console, but you couldn't isolate which pages were actually appearing inside AI-generated answers versus traditional blue links. Third-party tools offered estimates. Manual citation audits offered anecdotes. Neither offered the kind of first-party signal a CFO would accept in a pipeline review.
That changed on June 3, 2026, when Google launched dedicated Generative AI performance reports in Search Console. For the first time, you can see exactly which URLs Google's AI features are surfacing, how often, and where those impressions originate by country and device. The report covers AI Overviews and AI Mode in Search, with a separate view for generative features in Discover.
This is a diagnostic lens, not a scoreboard. Understanding what the report actually measures, and what it deliberately omits, determines whether you can turn the data into a decision.
What You're Actually Measuring
The new report tracks a single metric: impressions. According to Google's documentation, an impression counts each time a link to your site appears within a generative AI feature. You can slice that number by page, country, device, and date. That's the full inventory.
What the report does not include is the list that matters for revenue attribution: no queries, no clicks, no click-through rate, no average position, no indication of whether your page was cited as a source or merely displayed among links. Google says it is still working with website owners to determine which additional metrics would be useful. Translation: the report ships incomplete, and the roadmap is undefined.
This is not a minor omission. Seer Interactive's research found that organic CTR dropped 61% on queries where AI Overviews appeared, falling from 1.76% to 0.61%. But the same study found that brands cited in AI Overviews earned 35% more organic clicks than those that weren't. The difference between appeared and cited is the difference between brand awareness and pipeline contribution. The new report cannot distinguish between the two.
The Opt-Out Lever and Its Cost
Alongside the report, Google is testing a Search Console control that lets site owners exclude their content from AI Overviews, AI Mode, and generative Discover features. Opting out removes your links and content from grounding those features without affecting your traditional search rankings.
Before anyone reaches for that switch, the math deserves scrutiny. Ahrefs' December 2025 study of 300,000 keywords found that AI Overviews correlate with a 58% lower average CTR for the top-ranking page. That sounds like a reason to opt out. But the same research shows that position-one CTR on informational queries without AI Overviews has also declined, from 7.3% in December 2023 to 1.6% in December 2025. The erosion is structural, not isolated to AI features.
Opting out forfeits whatever brand visibility and citation traffic AI features might send, without restoring the CTR environment of 2023. The report can help you establish a baseline before making that call. If your AI impression volume is trivial, the decision is low-stakes. If it's substantial, you're trading a known quantity for an unknown recovery.
Reading the Report Without Query Data
The absence of query data is the sharpest limitation. You can see that a page appeared in AI features 50,000 times last month, but you cannot see which questions triggered those appearances. That makes it difficult to reverse-engineer what Google's AI considers your page authoritative for.
One practical workaround uses a regex filter in the standard Performance report. By filtering for queries nine or more words long, you can approximate the conversational, multi-clause prompts that tend to trigger AI Mode. The filter won't capture every AI-influenced query, but it gives you a proxy for AI-style search behavior while you wait for Google to add query data to the dedicated report.
The workaround is imperfect. AI systems often reformulate user prompts before running searches, so the query you see in Search Console may not match the question the user actually asked. But imperfect signal beats no signal when you're trying to understand which content themes are earning AI visibility.
What the Impression Number Actually Tells You
An AI impression confirms that Google's generative system found your page relevant enough to surface. It does not confirm that the user noticed your link, that your content was quoted in the answer, or that the appearance influenced any downstream behavior. The report answers did I appear? and almost nothing else.
That narrow answer still has value. If a page you expected to perform in AI features shows zero impressions, you have a content relevance problem or an indexing problem. If a page you didn't expect shows high impressions, you've discovered a topic where Google's AI considers you authoritative, which is worth understanding even if you can't yet measure the traffic it sends.
The report also lets you track trends. If your AI impressions are climbing while your traditional organic clicks are falling, you're watching the shift in real time. If both are falling, your content may be losing relevance across both surfaces. The directional signal matters for resource allocation, even without click data.
Connecting Impressions to Pipeline
The gap between AI impressions and revenue is wide, but not unbridgeable. Start by cross-referencing the pages with high AI impressions against your CRM's content-assisted opportunity data. If a page that appears frequently in AI features also shows up in the content journey of closed-won deals, you have a plausible attribution path. If it doesn't, the impressions may be brand awareness at best.
For pages where AI impressions are high but CRM attribution is absent, run a simple test: add a unique UTM parameter to the canonical URL and monitor whether any traffic arrives with that tag. This won't capture users who see your link in an AI Overview but don't click, but it will tell you whether the appearances are generating any measurable visits at all.
The harder question is whether AI visibility that doesn't generate clicks still contributes to pipeline. Seer's April 2026 update suggests the CTR decline has stabilized, which may indicate users are adapting to the new SERP layout. If clicks from AI features eventually recover, today's impression baseline becomes tomorrow's traffic forecast. If they don't, you're measuring brand exposure, not demand generation.
A Two-Week Pilot Plan
Before building dashboards or reallocating budget, run a contained test. Export your AI impression data by page for the last 30 days. Identify the top 10 pages by AI impression volume. For each, pull the corresponding data from your standard Performance report: clicks, CTR, average position. Then pull CRM data on content-assisted opportunities for those same URLs.
You're looking for three patterns:
- Pages with high AI impressions and high CRM attribution are your current AI-to-pipeline performers; protect and expand them.
- Pages with high AI impressions but no CRM attribution are brand awareness assets; decide whether that's worth the content investment.
- Pages with low AI impressions but high CRM attribution are your traditional organic performers; monitor whether AI features are cannibalizing their traffic.
The pilot won't answer every question, but it will tell you whether the new report is diagnostic noise or a signal worth operationalizing. Model or it didn't happen.
