Google's AI Overviews now reach about 2.5 billion monthly users — and on the queries where they appear, most searches end without a click. Here's the researched, no-hype read on the zero-click shift: the real CTR data, why being citednow beats ranking, and why monitoring AI answers needs real mobile-IP SERP tracking across locations.
AI Overviews reach ~2.5B monthly users and appear on roughly a quarter of US searches (near-universal on informational queries). On those queries clicks collapse — 83% end without a click (Semrush), and only 1% click a link inside the overview (Pew). The game shifts from ranking #1 to being cited inside the answer. And because AI answers are localized and device-specific, monitoring them needs real mobile IPs across geographies — the same demand-side of the closing web.
At I/O 2026 Google put AI Overviews at roughly 2.5 billion monthly users. Coverage has climbed fast: Conductor measured AI Overviews on about 25% of Google searches (November 2025), and Ahrefs found 99.9% of informational keywords now trigger one. The fuller conversational experience, AI Mode, processes over 1 billion queries a month.
The distribution matters: AI Overviews dominate informational intent and are far rarer on transactional or e-commerce queries (one January-2026 dataset put e-commerce near 4%). So the click loss is concentrated where the user just wants an answer — exactly the queries publishers relied on for traffic.
Different studies use different methods, so treat these as a range rather than one figure — but every credible source points the same way on informational queries:
Across ~68,000 queries: 8% click rate with an AI Overview vs 15% without — a 46.7% relative decline. Just 1% of users clicked a link inside the overview.
83% of AI-Overview searches end without a click; 93% in AI Mode.
~58% lower average CTR for the top-ranking page when an AI Overview is present.
This is the demand-side mirror of the supply-side story in The Closing Web: publishers are squeezed on one end by AI crawlers taking content, and on the other by AI answers taking the click.
When the answer is synthesized at the top of the page, position-1 is no longer the finish line. The click that doeshappen disproportionately goes to the sources the AI Overview names and links. Multiple 2026 analyses found brands cited in AI Overviews earn materially more clicks than uncited pages, and branded queries can even see CTR rise when a clear branded answer surfaces with a recognizable source.
The strategic flip: optimize to be quotable and attributable — structured, factual, schema-marked content an AI Overview can cite — rather than chasing a blue-link rank that increasingly goes unclicked. This is the heart of answer-engine optimization (AEO) / generative-engine optimization (GEO).
You can't optimize for what you can't see — and AI Overviews are not uniform. The overview a user gets is shaped by location, device, language, and the requesting network. The answer in Dallas on mobile differs from Berlin on desktop. So measuring "are we cited, and what does the AI say about us and our competitors?" at scale requires querying from real, geographically diverse, mobile network identities.
A datacenter IP gets a stripped, bot-flagged, or non-localized result — useless for AEO measurement. Real mobile-IP SERP monitoring across countries and carriers is what lets you track AI-answer presence accurately, the same way you must look like a real visitor to pass the detection stack.
The Coronium angle: our AI Visibility Lab and SEO-monitoring proxies exist for exactly this — tracking AI Overviews, AI Mode answers, and local SERPs from authentic 4G/5G devices in 20+ countries.
Track AI Overviews, AI Mode & local SERPs from real mobile devices.
AEO/GEO audits — measure your presence inside AI answers.
The supply-side mirror: AI crawler blocking & Pay-Per-Crawl.
Publishers squeezed by crawlers taking content and answers taking clicks.
Why monitoring needs a real-visitor network identity.
The regulatory layer of the AI data shift.