Google Tests AI Search Reports in Search Console

Google has quietly begun testing two new features in Search Console, giving select website owners more control and insight into how their content performs in the company’s generative AI Search experiences. One feature is a toggle that lets publishers opt their site out of appearing in AI Overviews, AI Mode, and AI Overviews within Discover. The other is a dedicated performance report that tracks how often URLs from a site are surfaced in these same AI-driven features across both Search and Discover.

Initially, these tools are rolling out only to a subset of UK-based websites. Google plans to expand availability to a global audience after the initial testing phase concludes.

The new AI visibility toggle is a significant step for site owners who want to manage their presence in live generative AI results. If a site uses this control to exclude itself, it will not receive traffic or impressions from AI Overviews, AI Mode, or AI Overviews in Discover. Importantly, Google has stated that this toggle will not influence a site’s ranking in traditional search results outside of these specific AI features. The company frames this as a natural extension of existing tools like snippet controls and Google-Extended. While snippet controls manage how content appears in standard search snippets and Google-Extended blocks content from training AI models, this new toggle is exclusively about appearing in the live, user-facing generative AI Search experiences.

The accompanying new performance reports offer a dedicated lens into AI visibility. According to Google’s announcement, these reports show how often a site’s URLs were displayed in generative AI features on Search and Discover. The data can be broken down by page, country, device, and even down to the hourly level. While generative AI impression data has been included in the overall Search Console performance report for some time, this new view isolates that AI-specific data for clearer analysis.

A notable omission from these reports is click data and query-level metrics, which are critical for site owners trying to understand actual user engagement. Google acknowledges this gap, stating it is “continuing to work with website owners to understand what insights will be most helpful” and plans to add more metrics over time.

This development comes after more than a year of requests from site owners for AI-specific data in Search Console, dating back to the U. S. launch of AI Overviews in 2024. When Google confirmed that AI Mode data was counted in overall Search Console totals, there was no way to separate AI-driven traffic from organic search results. Google’s John Mueller previously clarified that all links within an AI Overview share a single position in Search Console, making it difficult to assess which placements drive value. Our reporting has consistently highlighted this measurement gap. Just two weeks ago, we noted that Google had expanded link surfaces in AI features without providing click data for those surfaces.

Meanwhile, Microsoft has moved more quickly on this front. Bing Webmaster Tools launched an AI Performance dashboard in February, added grounding query-to-page mapping in March, and previewed a Citation Share feature at SEO Week in May.

For site owners, these new reports offer the first dedicated view of a website’s visibility in Google’s AI Search features. This is a clearer picture than previous bundled reports, where AI-driven visibility was hard to isolate. However, the missing piece remains click data specific to AI features. Impressions tell you how often your pages appeared, but not how often people clicked through from those appearances. That gap has been the central question in our coverage of AI Search measurement for over a year, and this announcement doesn’t close it yet.

Looking ahead, Google says it will test both features with a subset of UK websites before expanding them globally. The company has committed to adding more metrics to the reports over time, but has not named specific additions or timelines.