The 58% problem: what AI search did to your clicks in 2026

When Google answers the question itself, the page ranking first loses more than half its clicks. Ahrefs measured it across 300,000 keywords: by December 2025, the presence of an AI Overview correlated with a 58% lower click-through rate (CTR) for the top result, up from 34.5% nine months earlier. The click, the unit this industry priced everything in, is losing its place as the default outcome of a search.

But the same data shows a smaller, richer market underneath. The traffic that survives behaves differently, and brands cited inside AI answers are pulling away from everyone else. This article walks through the verified numbers: how far clicks fell, and what measurably earns visibility now that engines answer first and link second.

How steep the fall was Pew Research Center watched real behavior. In a 2025 study spanning 68,879 Google searches from 900 US adults, people clicked a traditional result on 8% of visits when an AI summary appeared, against 15% when it didn't. And links inside the summary did almost nothing for publishers. Users clicked those on 1% of visits.

Agency data agrees. Seer Interactive tracked 3,119 search terms across 42 client organizations and watched organic CTR on AI Overview queries fall from 1.76% in June 2024 to 0.61% by September 2025, a 61% decline. Paid ads on those queries did worse, dropping 68%.

The damage isn't evenly spread, though. Amsive's study of 700,000 keywords found non-branded terms lost 19.98% of their CTR, while branded keywords that triggered AI Overviews gained 18.68% more clicks. If people search your name, the AI answer works like a billboard.

Google says you're measuring it wrong

So who's right about the damage? Google's position is blunt. Liz Reid, vice president and Head of Google Search, wrote in August 2025 that "total organic click volume from Google Search to websites has been relatively stable year-over-year" and that Google sends "slightly more quality clicks" than a year before. On the Q2 2026 earnings call, CEO Sundar Pichai added that AI features in Search now send "billions of clicks to websites every week." Both things are true at once, because the pool itself exploded. AI Overviews passed 2 billion monthly users in July 2025, and AI Mode crossed 1 billion monthly actives after its global rollout last October. ChatGPT reached 900 million weekly users this February.

People ask more questions than ever. They just resolve most of them without visiting anyone.

Rand Fishkin's SparkToro analysis of Similarweb clickstream data puts a number on that habit: 68.01% of US searches in the first four months of 2026 ended without a click, up from 60.45% in 2024. Stable total clicks spread across a larger, hungrier query base still means each individual page gets less. Neither side's numbers are wrong. They're measuring different things, which is why publishers and the platform keep talking past each other.

Fewer visitors, better visitors

The traffic that survives the AI filter is worth more per visit. Semrush found the average AI search visitor is 4.4 times as valuable as a traditional organic visitor, measured across 500-plus digital marketing topics. Adobe Analytics, which drew on over 1 trillion visits to US retail sites, recorded 1,200% growth in generative AI referrals between July 2024 and February 2025. Those visitors behaved differently once they arrived:

· 8% higher engagement than other traffic sources

· 12% more pages browsed per visit

· 23% lower bounce rate

· 9% lower conversion rate, a gap that had been 43% seven months earlier

The premium for being the quoted source is sharper still. Seer Interactive's 2026 update, built on 5.47 million tracked queries, found brands cited inside an AI Overview earn 120% more organic clicks per impression than brands that aren't.

The same dataset caught the first recovery in the trend line: organic CTR on AI Overview queries climbed from a 1.3% floor in December 2025 to 2.4% by February 2026.

The market punished everyone, then started paying the cited. Nobody can promise that recovery holds. What the trend line does show is a market re-pricing visibility around citation, and re-pricings reward whoever measures first.

What earns a citation now

Not what the checklists promised. Ahrefs tracked 1,885 pages that added JSON-LD schema markup between August 2025 and March 2026, matched them against 4,000 control pages, and found no significant uplift in AI citations on any platform. I keep coming back to that result, because it kills the tidiest excuse in the industry: the belief that a technical fix would substitute for content worth quoting.

Content format is a different story. Wix's AI Search Lab analyzed 75,000 AI answers with over a million citations across ChatGPT, Google AI Mode, and Perplexity, and listicles took 21.9% of all citations, the largest share of any content type. On commercial queries their share reaches 40.86%. Put together, the studies argue for a short set of moves:

· Format key pages as structured, scannable breakdowns rather than essays

· Anchor every number to a named, linkable source

· Grow branded search demand, which AI answers amplify instead of eroding

· Measure citation share in AI answers, and drop schema from the priority list For a fuller working checklist, Snoika's breakdown of AI SEO best practices 2026 maps these findings onto page-level tactics, from answer placement to citation tracking.

The metric that replaces the click

Search didn't die. It stopped paying in the currency everyone budgeted for. The behavioral studies and the conversion data describe the same market: a smaller stream of visitors who arrive with higher intent, and a widening gap between brands that AI engines quote by name and brands they merely summarize.

So change what you measure. Track citations the way you tracked rankings, and build pages an engine can lift a whole answer from. The vanished clicks aren't returning. But the visitors who still click were always the ones worth having.

Snoika OU
Harju maakond, Tallinn, Pirita linnaosa, Tohu tn 10-7, 12013
Anton Vedeshin, PhD. Founder of Snoika

Snoika is an AI-first visibility and growth platform that helps businesses become found, cited, trusted, and recommended across AI search engines and traditional search. The platform measures how brands appear in AI-generated answers from tools such as ChatGPT, Gemini, Claude, and Perplexity, identifies visibility gaps against competitors, and helps companies improve their presence through AI search optimization (GEO/AIO), SEO, content, digital PR, and other authority-building channels.

This release was published on openPR.