Google finally let you measure AI search

What actually shipped on 3 June

Google launched dedicated Search Generative AI performance reports in Search Console, its first native tooling for seeing how a site shows up inside AI Overviews and AI Mode. Until then, impressions from AI features were folded into general Search data with no way to isolate them. An industry roundup described it plainly as the first native measurement of AI Overviews and AI Mode impressions.

Google also announced compliance with a UK regulatory requirement that it offer publishers a way to opt out of being aggregated into AI search. The Competition and Markets Authority called the move a world first. The opt-out toggle itself became effective on 17 June, and it removes content from AI features without touching organic rankings.

The pressure had been building. The CMA had first designated Google as having strategic market status in October 2024, then pushed it in January 2026 to give publishers a choice over whether their content feeds AI features. June was when that produced a shipped control.

The click was already leaving

The report lands at a point where most searches no longer end in a click. Rand Fishkin's SparkToro analysis of a clickstream panel found that 68.01% of US Google searches ended without a click between January and April 2026, up from 60.45% in 2024. Similarweb, looking at news specifically, found zero-click news searches on Google rising from 56% when AI Overviews launched in May 2024 to almost 69% by May 2025. The two studies count different things (SparkToro all searches, Similarweb news only), but both show fewer searches sending anyone onward.

BrightEdge's one-year review of AI Overviews reported that they now appear in over 11% of Google queries, with a near-30% reduction in click-through rate since May 2024. That figure does not stand alone. Ahrefs, studying 300,000 keywords independently, found AI Overview presence cut position-one click-through by roughly 34.5%. Amsive, studying 700,000 keywords across ten sites in five industries, found that the roughly 10,000 keywords triggering an AI Overview saw an average click-through drop of 15.49%. The Ahrefs study is from April 2025, not 2026; it keeps recurring in this year's reports because the effect it measured has not gone away. When an AI answer sits above your link, far fewer people click through. What now matters is whether you were cited at all.

Ranking no longer buys you a citation

A top-10 ranking used to strongly predict a citation. That link has weakened sharply. Ahrefs analysed 863,000 keywords and 4 million AI Overview URLs and found that only 38% of pages cited in AI Overviews also appeared in the top 10 for the same query, down from 76% in the July 2025 version of the study.

One caveat: Ahrefs improved its parsing method between the two studies, so the datasets are not directly comparable. Treat the fall from 76% to 38% as a strong signal of direction rather than a precise measurement of collapse. Either way, rank and citation are now separate things to check. A page can rank first and still be invisible to the AI answer.

The engines also disagree on who they cite. Running 100,000 identical prompts through both ChatGPT and Perplexity, Profound found that only 11% of cited domains appeared in both engines - 37.4% in ChatGPT only, 51.6% in Perplexity only. A separate study by Writesonic, across 161,286 prompts on four platforms, reported a roughly 13% overlap between ChatGPT and Perplexity, with only 3.8% of citations shared across all four. The mechanism is a difference in source diet: across 680 million citations Profound tracked, Wikipedia made up 47.9% of ChatGPT's top-ten source share while Reddit ran to 46.7% of Perplexity's. Semrush's index put a sharper edge on the wider point: analysing 126 million US prompts from January to April 2026 across ChatGPT, Gemini, Google AI Mode and Google AI Overviews, it found that only 36 brands held top-100 visibility across every platform in every month of the study.

AI answers can also work against you

Being visible is one thing. What the AI says about you is another, and it can be negative. BrightEdge reported that Google AI Overviews are 44% more likely than ChatGPT to surface negative brand sentiment, carrying it in about 2.3% of brand mentions versus roughly 1.6% for ChatGPT. The same figures were picked up by Fortune. Those are small percentages of a very large surface: BrightEdge notes that over three billion people interact monthly with Google AI Overviews and ChatGPT combined.

This is where the June opt-out earns its place. A publisher whose content is being misrepresented can now pull it from AI features without losing organic rankings. It is a blunt tool, but a real one.

Most teams still cannot see this

The arrival of Search Console AI reports matters because the measurement gap has been wide. In Semrush's survey, 45% of marketing leaders said they cannot accurately measure their brand's visibility in AI-generated answers, even as that channel grows.

The same study found the payoff for closing that gap. Among organisations that fully integrate their SEO and AI-visibility work, 81% reported increased traffic or leads from AI platforms. This is survey data, so read it as correlation: the teams that measured reported better results than the ones guessing.

A note on reading the numbers in this piece. Several come from SEO platforms reporting on their own tooling, and single studies are stronger on direction than on precision. Where a figure comes from a vendor or a survey, treat it as a directional signal. The methods differ, but the findings line up.

Turn the new visibility into a routine

Four things worth doing this month:

Measuring the answer layer

Search Console now shows your AI impressions on Google. The harder question spans ChatGPT, Gemini and Perplexity too: are you cited at all, and what does the answer say? That is what we build for.

Because the measurement is new, an early baseline is worth having. Set up tracking across ChatGPT, Gemini and Google today.

Measure your AI visibility