How to Read Current Google AI Search Reporting Alongside Ordinary Search Data

Google Search Console now has a dedicated report for impressions from links to your site in AI Overviews and AI Mode, Google Search’s AI features. It also has a broader Search performance report. An impression means Google recorded a link to your site as shown; it does not tell you whether someone clicked, visited your site, or became a customer.

Read the two reports together, but do not add their totals. AI-feature data is included in the ordinary Search report’s Web totals. The dedicated report helps you see where Google recorded AI-feature impressions, while the Web report shows overall activity through clicks and impressions, along with click-through rate and average position. Those broader metrics do not isolate AI-feature activity.

Google announced the dedicated report in June 2026 and said it was available worldwide by August 31, 2026. The practical question is how to compare its data with ordinary Search results without treating related changes as proof that one caused the other.

What each Search Console report measures

The Generative AI performance report (Search) records impressions for links to your site in Google Search’s AI Overviews and AI Mode. It groups results by page, country, date, and device. Its search-type filter also separates text-based Web searches from searches that use more than text, such as searches that begin with an image. The report does not provide AI-specific clicks, queries, click-through rate, or average position, according to Google’s report documentation.

The ordinary Performance report, with search type set to Web, covers a wider set of Search results. It reports overall clicks and impressions, along with click-through rate and average position; its dimensions include queries and pages. Google includes AI Overview and AI Mode activity in this Web reporting. A click on an external link in an AI feature counts as a click, but the general report does not separate those clicks from other Web clicks. Google describes these reporting rules in its explanation of clicks, impressions, and position.

Use the AI report to see where Google recorded impressions from AI features, and the Web report to track the site’s overall Search performance. Because the reports overlap, adding AI impressions to Web impressions counts some exposure twice.

Compare the reports without claiming attribution

To compare the reports, select the same Search Console property and date range in both. Start with the AI report’s page and date views. Then open the Web Performance report for that property and period to review overall clicks and impressions. If a page’s AI impressions rise while its Web clicks fall, note both changes and investigate. The reports alone do not show that one caused the other.

This distinction matters because the dedicated AI report does not report AI traffic. It cannot show whether a particular AI impression led to a click, or which query prompted Google to show the link. The Web report’s query table provides query data for overall Search, but it cannot identify which queries produced AI-feature impressions. Google also withholds some queries for privacy and limits the rows shown, as described in its documentation on Search Console dimensions and data groupings.

Read average position with care. It summarizes where a site’s top result appeared across reported searches, but it does not show the exact layout any one person saw. Google assigns links in an AI Overview the position occupied by the overview, so Web average position is not a clean measure of ordinary blue-link rankings.

Check your own property

Open the Search Console property that covers the site or URL you want to assess. Use the same property in both reports, especially if you use both domain and URL-prefix properties.

Then:

1. Open Generative AI performance report (Search). Review impressions by page, country, date, and device. Check text-based and multimodal Web search where each applies to your site.
2. Select that same property and date range in the Search results Performance report. Leave search type set to Web, then review clicks, impressions, click-through rate, and average position. Use page and query views to investigate overall Search trends.
3. Compare the patterns, not the totals. See whether pages with recorded AI impressions also show changes in overall Web performance, then check visits and conversions for those pages in your analytics. Treat these as changes that happened at the same time unless you have separate evidence that shows one caused the other.

Google’s Search Console and Analytics integration can put Search Console data alongside Analytics landing-page metrics. This helps you see what happened after Search visits, but it does not show whether AI Overviews or AI Mode specifically brought those visits or conversions.

For comparisons over time, record which property and date range you used, along with the search type and filters. The AI report’s recent data can be preliminary. Its chart and page table can also show different totals: the chart counts at the property level, while the table groups by page. If multiple results from the same site appear in an AI feature, Google counts one impression in the chart. Google also generally assigns performance data to a canonical URL, the address it treats as the main one for a page. Because of these reporting rules, different totals in the chart and table do not automatically mean there is a tracking problem.

If the AI report is missing or a page has no impressions

An absent report does not by itself show that a page has a technical error. Google says the report may not appear when a site has not received enough AI-feature impressions. A setting for Search generative AI can also keep a property out of these features. Check the Search generative AI control, including whether the property inherits a setting from a parent property.

If the report appears but a page has no impressions, check whether Google has indexed the page, meaning added it to its Search index, and whether Google can show it with a snippet, the short description that appears for a page in Search. Google says pages must meet these conditions to be eligible to appear as links supporting AI Overviews or AI Mode. Use URL Inspection and Search Console’s indexing reports to investigate. Eligibility does not guarantee that Google will select the page for an AI feature. Even if Google successfully accesses the page, that does not prove Google indexed or displayed it. Google’s AI features guidance explains how eligibility relates to existing Search requirements.

Investigate a sudden change in impressions before deciding what caused it. Check that the date range, country, device, and search-type filters match your comparison. Then look for relevant changes to the page and compare the AI report with overall Web impressions and clicks. These checks can show what changed in the recorded data, but the cause remains undetermined by these checks.

What the numbers cannot tell you

The dedicated report leaves three details unknown: the exact wording behind an AI response, an AI-specific click-through rate, and a ranking just for AI features. The ordinary query report covers Web Search overall, so it cannot supply those details. Do not infer that an AI impression represents a specific question, a recommendation, or a visit.

An impression means Google recorded a link as displayed under its reporting rules. It does not show whether a searcher noticed the link, whether an AI response used the page accurately, or whether seeing the link influenced a later purchase. Analytics can help you assess visits and conversions on your site, but the combined reports do not prove that AI visibility caused those outcomes.

Google AI features can also explore related subtopics when preparing a response, and AI Overviews do not appear for every query. So one search you run by hand does not represent the reporting data. Use Search Console to understand Google’s recorded activity, and use other evidence for questions the reports do not answer.

Key Takeaways

• Keep the reports separate but read them together: AI-feature impressions are included in overall Web reporting, so adding the totals double-counts some exposure.
• Use the AI report for recorded visibility: It groups impressions by page, country, date, and device, but does not report AI-specific clicks, queries, click-through rate, or position.
• Compare matching properties and dates: Treat changes in AI impressions and Web clicks as patterns to investigate, not proof that one caused the other.
• Check settings and eligibility before diagnosing absence: Insufficient AI impressions, an exclusion setting, and page indexing or snippet eligibility are separate possibilities.
• Use analytics for on-site outcomes with care: Landing-page visits and conversions add context, but do not establish that AI features caused them.

What I'm building

Delegate tasks. Get software.

Give Vroni a GitHub issue, bug report, spec, or rough idea. It reads the repo, plans the change, writes code, runs checks, and works toward a review-ready pull request.

Take a look at vroni.com