Search Console doesn't separate AI Overviews clicks, and there's no filter for that

Search Console doesn't separate AI Overviews clicks, and there's no filter for that

Editor's note: the text below mentions that the Google Search Central documentation was updated on December 10, 2025; the correct date recorded on the official page and in this outlet's previous coverage is December 31, 2025. We kept the rest of the content, which addresses the lack of click segmentation for AI Overviews in the Performance report, since this is the article's central point and is correctly supported by the source.

Google Search Central documentation confirms that AI Overviews and AI Mode fall under the 'Web' search type in the Performance report, without their own segmentation. The viable approach is to observe aggregate trends and cross-reference with Analytics, not isolate the click.

Updated Google Search Central documentation confirms that AI Overviews and AI Mode continue to be counted under the 'Web' search type in the Performance report. There is no search appearance filter with a dedicated option for AI Overviews: what can be done is to observe the variation in aggregate traffic over time and cross-reference it with Analytics data, without isolating the click that came specifically from an AI-generated answer.

The narrative circulating among growth teams usually starts from a premise worth checking before moving any budget: that Search Console would already have a dedicated counter to separate, line by line, what came from AI Overviews from what came from traditional search. Google Search Central's documentation on AI features, updated on December 10, 2025, says the opposite, to the letter. According to the text, pages that appear in AI Overviews or AI Mode "are included in Search Console's overall search traffic" and are reported within the Performance report, under the same 'Web' search type used for classic results. There is no dedicated search type for AI within the Performance report.

This isn't a bureaucratic detail. If the growth team is trying to justify investment in content optimized to appear inside AI-generated answers, the first thing to align with whoever is demanding the number is that Search Console, by default, doesn't return a click or an impression marked as "this came from AI Overviews." What it returns is the sum of everything under 'Web', and there is no native segmentation to separate out that fraction, as this same outlet has previously reported.

Why the click can't be isolated

The official documentation doesn't describe any search appearance filter with a specific option for AI Overview in the Performance report. The text is direct: sites that appear in AI features "are included in overall search traffic" and reported under the 'Web' type, without their own segmentation. This confirms what was already known: there is currently no native way in Search Console to filter clicks, impressions, CTR, or average position specifically for appearances in AI Overviews or AI Mode.

The AI features documentation reinforces a point that helps interpret this limitation more generously than it usually gets: "we've seen that when people click from search results pages with AI Overviews, these clicks are higher quality," meaning Google itself states that clicks originating from result pages with AI Overviews tend to translate into more time spent on the site afterward. This is an argument in favor of measuring engagement depth (via Google Analytics, as the documentation itself recommends cross-referencing) rather than just click volume, when the question is whether it's worth continuing to invest in GEO for a given content cluster, even without being able to isolate the exact share of traffic coming from AI Overviews.

What gets mixed together under 'Web'

Here's the point that any growth analysis needs to record as a limitation, not as a minor detail. Without native segmentation, a click reported in the Performance report may have come from inside an AI Overview box, from a classic result right below it, or from any other element on the same results page: the report doesn't distinguish the exact origin within the SERP layout. Search Console's click count has always been aggregated by URL and by query, never by position of origin, and that logic doesn't change for AI Overviews.

Another gap that matters for anyone calculating GEO ROI: without an isolated metric, an aggregate CTR variation for a given content cluster could be reflecting either a change in click behavior within AI-generated answers or a drop in content relevance in classic results, and the report by itself doesn't separate the two causes.

It's also worth noting that the documentation is explicit that there is no additional technical requirement to appear in AI Overviews or AI Mode beyond what already applies to classic search: an indexed page eligible to display a snippet. The SEO fundamentals cited in the material (allowing crawling in robots.txt, internal links that make content discoverable, good page experience, available text content, quality images and videos supporting the text, structured data consistent with the visible content, and an up-to-date Merchant Center or business profile listing) are the same ones that underpin any technical SEO strategy. There is no parallel GEO checklist with its own crawling rules; controlling Googlebot's access via robots.txt remains the site owner's lever for both traditional search and the AI layers built on top of it.

How this changes growth work

In practice, a reasonable approach for growth teams that need to present an honest reading, rather than an exact figure the product doesn't deliver, is: track the variation in total traffic (the 'Web' type) for a content cluster over time, observing windows before and after relevant SERP changes for those queries, and cross-reference that variation with session-time and conversion data in Google Analytics. This doesn't isolate the exact share of traffic coming from AI Overviews, but it helps form a hypothesis about the direction of the effect, especially when cross-referenced with Google's own claim about click quality, validated (or refuted) for the specific case of the site in question, rather than accepting the company's published average as a universal truth for any domain.

A point of attention for anyone proposing an opt-out: noindex, nosnippet, data-nosnippet, and max-snippet remain the correct controls to limit what appears in AI features, exactly as with classic snippets. This is different from Google-Extended, which the documentation itself separates out as a control aimed at training and grounding in other Google systems, not the AI features within Search. Confusing the two leads to blocking decisions that don't produce the expected effect.

What still remains open

Google doesn't disclose, either in Search Console or in the AI features documentation, a metric that separates a click from inside an AI Overview box from a click outside it on the same SERP, nor a native filter to isolate that share of traffic. Until that changes, the most honest reading a growth team can deliver is the aggregate variation of 'Web' traffic over weeks, cross-referenced with post-click engagement in Analytics, enough to separate trend from noise, but not enough to produce an exact figure of how much came from AI Overviews. Treating that number as a directional estimate, not a definitive metric, is what prevents budget decisions based on a precision that Search Console itself still doesn't deliver.