Google Search Console Generative AI ReportWhat It Shows and How to Use It

Summary

Google added a generative AI performance report to Search Console in June 2026 and rolled it out worldwide by the end of August. It counts how often your URLs appear in AI Overviews, AI Mode and AI features in Discover, by page, country, device and date. It shows no clicks, no queries and no answer text. On getrefine.ai, 509 AI impressions over 90 days revealed a pattern classic reports hide: comparison, listicle and platform pages over-index in AI answers. This guide explains how to read the report and what to pair it with.

What Is the Search Console Generative AI Report?

The generative AI report is a dedicated view in Google Search Console that isolates impressions earned in Google’s AI surfaces. Google announced it on June 3, 2026, first for a subset of UK sites, then for all properties (Google Search Central). You find it in the Performance section, under a link labelled for generative AI in search results. Data starts on May 18, 2026, with no backfill.

Before this report, AI Overviews and AI Mode impressions were blended into the regular Web totals. You could not tell whether a page was gaining exposure in AI answers or in blue links. That separation is the whole value of the report.

What the Report Measures (and What It Leaves Out)

The report gives one metric, impressions, and four dimensions:

• Pages: which of your URLs appeared in an AI feature.
• Countries: where those impressions happened.
• Devices: desktop, mobile or tablet (for Search).
• Dates: daily, weekly or monthly trend.

What it leaves out matters as much:

• No queries. You do not see the question that triggered the AI answer.
• No clicks or CTR. AI-driven clicks stay mixed into the classic Web report.
• No answer content. You cannot tell if you were recommended, merely linked, or described inaccurately.
• No competitors. You only see your own property.
• No mentions without links. If Google names your brand but cites another site, nothing shows up for you.

Our Own Data: Which Pages Google Surfaces in AI Answers

Here is what the report showed for getrefine.ai between June 29 and September 28, 2026: 509 AI impressions in total, against about 11,000 classic search impressions over the same window. The interesting part is not the total but the ratio per page.

• Homepage: 212 AI impressions for 1,832 classic impressions (about 12%).
• Claude platform page: 30 AI impressions for 92 classic (about 33%).
• “Refine vs Semrush and Ahrefs” comparison: 16 AI for 98 classic (about 16%).
• “Best AI visibility tracking tools 2026” listicle: 44 AI for 335 classic (about 13%).
• Blog index: 17 AI for 862 classic (about 2%).
• SEO professionals use-case page: 17 AI for 535 classic (about 3%).

Two lessons. First, pages that answer a specific evaluation question (a comparison, a shortlist, a platform-specific page) are over-represented in AI answers relative to their classic exposure. Second, hub pages and broad landing pages earn plenty of blue-link impressions but rarely get pulled into AI answers. These are small numbers from one site over 90 days, so treat them as a direction rather than a law, but the pattern matches what we see when tracking AI answers directly.

How to Read the Report in 15 Minutes a Week

A weekly routine that takes about 15 minutes:

• Check the trend on the Dates tab. A sudden drop across all pages usually means an AI feature change; a drop on one page means that page lost ground.
• Review the top 10 pages. Note new entries and exits since last week.
• Compare with the classic report. Pages with high classic impressions and zero AI impressions are refresh candidates: add a direct answer at the top, a table, and the sub-questions buyers ask.
• Annotate changes. Log every content update with its date so you can compare four weeks before and after.
• Check countries if you sell in several markets. AI features roll out unevenly.

If you want the AI Mode angle specifically, our guide on tracking brand mentions in Google AI Mode goes deeper. For the traffic side, see how to track AI traffic in GA4.

Pair It With Bing’s AI Performance Report

Microsoft’s equivalent is more generous on one point. Since February 2026, the AI Performance report in Bing Webmaster Tools shows citations of your pages in Copilot and Bing AI answers, and the grounding queries Copilot generated to retrieve them (Bing Webmaster Tools). Grounding queries are not what the user typed, but they show how an AI engine rewrites questions in your category, which makes them a free source of prompts to track.

Used together, the two reports give you a first-party baseline of AI exposure on the two largest search indexes. Bing reports a sample rather than complete activity, so read trends, not absolute numbers.

Where Search Console Stops and Prompt Tracking Starts

Search Console answers “was my page shown in an AI feature?”. Buyers and leadership care about a different question: “when someone asks AI for options in our category, are we recommended, and over whom?”. That requires running a fixed set of prompts on a schedule and reading the answers: mention rate, position, sentiment, sources cited, and share of voice against competitors.

Treat the generative AI report as your early-warning system and your page-level map. Treat prompt tracking as your scoreboard. You need both to know whether GEO work is paying off.