The state of AI search (2026): citation, zero-click, and referral benchmarks
The short version: Google searches now end without a click 68 percent of the time, AI Overviews appear on roughly half of tracked queries, and only 38 percent of their citations still come from the traditional top ten (SparkToro, Ahrefs, both to re-verify). Meanwhile the traffic AI engines do send converts at multiples of ordinary organic traffic. Fewer clicks, higher-value clicks: that is the shift a brand is actually navigating.
Every figure below is compiled from published, third-party research, not from Thabrew Effect's own data. It is presented as an article, not an original study, and each number carries its source so it can be checked rather than taken on faith.
Why does zero-click search matter more than ever?
For most of a search's twenty-year history, a click was the unit everyone measured. That unit is disappearing. In the first four months of 2026, 68.01 percent of Google searches in the US ended without a click, up from 60.45 percent in 2024 and 49 percent back in 2019, according to a SparkToro analysis of Similarweb clickstream data (SparkToro, to re-verify). A separate Bain and Company estimate puts the baseline no-click rate on traditional search at around 60 percent, and Pew Research found only 8 percent of users click a traditional link at all when an AI summary appears on the page (Semrush, citing Bain and Pew, to re-verify). The reader has not gone away. The reader has simply stopped needing to click to get an answer.
How often does an AI Overview show up, and does ranking still earn the citation?
By March 2026, AI Overviews were appearing on roughly 48 percent of tracked queries across nine commercial industries, up 58 percent from a year earlier, per BrightEdge data reported by Search Engine Journal (Search Engine Journal, to re-verify). What happens inside that box has changed just as fast. Ahrefs studied 4 million AI Overview citations across 863,000 keyword searches and found only about 38 percent came from the traditional top ten results, down sharply from roughly 76 percent when Ahrefs ran the same analysis in July 2025 (Ahrefs, to re-verify). A page ranking well is no longer a reliable proxy for a page getting quoted.
68%
of Google searches end without a click
48%
of tracked queries now show an AI Overview
38%
of AI Overview citations come from the top 10, down from 76% a year earlier
Which engines are actually sending the traffic?
ChatGPT remains the dominant source of AI referrals by a wide margin. Across the ten industries in Conductor's 2026 AEO/GEO Benchmarks Report (13,770 domains, 3.3 billion sessions analyzed), ChatGPT accounted for an average of 87.4 percent of all AI referral traffic (Conductor, to re-verify). Separately, Similarweb-sourced market share data reported in January 2026 put ChatGPT at 68 percent of AI chatbot web visits worldwide, ahead of Gemini at 18.2 percent, DeepSeek at 4 percent, Grok at 2.9 percent, and Perplexity and Claude near 2 percent each (Vertu, citing Similarweb, secondary, to re-verify). Shares like these move fast. Both Gemini and Claude have posted year-over-year growth in the hundreds of percent even while ChatGPT holds the larger base, so a snapshot from one month can read differently by the next.
Do AI-referred visitors actually convert better?
This is the number every brand asks about first, and the honest answer is that it depends which study is in front of you. Conductor's July 2026 report found AI-referred visitors converting at roughly twice the rate of other traffic sources, in about a third of the sessions (Conductor, to re-verify). Semrush put the average AI search visitor's value at 4.4 times a traditional organic visitor (Semrush, to re-verify). Ahrefs, measuring its own site with Ahrefs Web Analytics, found AI search traffic was just 0.5 percent of visits but drove 12.1 percent of signups, a 23 times conversion differential (Ahrefs, a single-site result, to re-verify). Adobe Digital Insights, analyzing more than a trillion visits across 130-plus North American retailers, reported AI-referred shoppers converting 42 percent better than non-AI traffic in March 2026, up from converting 42 percent worse a year earlier (Adobe Digital Insights, via reporting, to re-verify). The multiples vary because the traffic is not the same kind of visitor. Someone arriving because ChatGPT named a brand by name is closer to a referral from a trusted advisor than a curious click from a results page.
The benchmark table
| Metric | Figure | Source | Date |
|---|---|---|---|
| Zero-click rate, all Google searches (US) | 68.01% | SparkToro / Similarweb | Jan to Apr 2026 |
| AI Overview appearance rate (9-industry tracker) | ~48% of tracked queries | BrightEdge, via Search Engine Journal | Mar 2026 |
| AI Overview citations from the top 10 organic results | 38% (down from ~76% in Jul 2025) | Ahrefs | Mar 2026 |
| AI referral traffic as a share of total site traffic | 1.08% average (10 industries) | Conductor | Jul 2026 |
| ChatGPT's share of AI referral traffic | 87.4% | Conductor | Jul 2026 |
| AI chatbot market share by web visits (worldwide) | ChatGPT 68%, Gemini 18.2%, others under 5% each | Similarweb, via Vertu | Jan 2026 |
| AI-referred visitor conversion multiple | 2x to 23x, varies by study and site | Conductor, Semrush, Ahrefs | 2025 to 2026 |
| AI-referred retail shoppers converting better than non-AI traffic | +42% | Adobe Digital Insights | Mar 2026 |
| Share of AI citations from sources a brand does not own | 94% | Muck Rack | May 2026 |
| Share of ChatGPT citations that are Wikipedia | 7.8% | Profound | 2025 |
| Share of Perplexity responses citing Reddit | ~47% of responses | Agile Growth Labs, citing Profound | 2025 |
All figures above are to re-verify at source before publish, and several are single-study or secondary reporting rather than confirmed, repeated findings. Treat the wide range on conversion multiples as directional, not a single fact.
What does this mean for a brand's strategy?
Three things follow from the table, and none of them are exotic. First, the traffic that still exists after a zero-click search is worth chasing harder, not less, because it converts at a premium across every study measured here. Volume is down. Value per visit is up. Second, ranking is necessary but no longer sufficient. With only 38 percent of AI Overview citations still tracking the top ten, the content that gets named is the content structured to be lifted, not merely the content that ranks. Third, concentration is real and temporary. ChatGPT sends the overwhelming majority of AI referral traffic today, but Gemini and Claude are growing fast enough that a strategy built only for one engine is a strategy with an expiration date. A brand built for citation, not just ranking, is positioned for whichever engine wins the next round.
Frequently asked questions
About 68 percent, as of the first four months of 2026, according to a SparkToro analysis of Similarweb clickstream data. That is up from roughly 60 percent in 2024 and 49 percent in 2019, a steady multi-year climb.
Roughly 48 percent of tracked queries across nine commercial industries showed an AI Overview by March 2026, per BrightEdge data, up 58 percent from a year earlier. Prevalence varies by industry and query type.
Less than it used to. Ahrefs found only 38 percent of AI Overview citations came from the traditional top ten in March 2026, down from about 76 percent in July 2025. Ranking helps a page get considered, not necessarily cited.
ChatGPT, by a wide margin. Conductor's 2026 benchmark, drawn from 3.3 billion sessions across ten industries, found ChatGPT responsible for 87.4 percent of all AI referral traffic on average.
No, and the two numbers get confused often. Referral share measures who sends clicks to other sites. Market share measures who gets visited directly. Similarweb-sourced data put ChatGPT's worldwide chatbot visit share at 68 percent in January 2026, with Gemini second.
Every study reviewed here says yes, though the size of the premium ranges widely: from roughly twice the conversion rate in Conductor's broad ten-industry sample to 23 times higher in a single-site Ahrefs analysis. Higher intent, not a trick of measurement, explains the gap.
Because they measure different traffic at different scales. A single site's AI referrals, an industry-wide sample, and a retail-specific dataset are not the same population. Treat the range, not any one multiple, as the finding.
No, not yet, and not close. Conductor's 2026 data put AI referral traffic at about 1.08 percent of total site traffic on average. It is a small, fast-growing, and disproportionately valuable slice, not a replacement.
A small one. Muck Rack reported that 94 percent of AI citations came from sources a brand did not own, which is why third-party coverage and community presence matter as much as the brand's own pages.
No. AI Overviews still draw from the same Google index as traditional search, and a page that is not properly indexed and ranked is rarely a page that gets cited either. The two disciplines have merged, not diverged.