AI Search Statistics 2026: Adoption, Share, Clicks and Traffic Value

Two years ago, "AI search" was mostly a thought experiment. In 2026 it is a measurable channel with its own traffic patterns, its own market share table and its own line in enterprise marketing budgets. What has changed is not just adoption. It is that we finally have enough independent measurement, from panel providers, regulators, analytics vendors and the platforms' own earnings calls, to say something useful about size and direction rather than vibes. The numbers below are drawn from primary sources published in 2025 and 2026. Where credible measurements disagree, and they frequently do, that disagreement is part of the story.

Key AI search statistics for 2026

• 49% of US adults now use AI chatbots, up from 33% in 2024, according to the Pew Research Center. [1]
• 54% of UK adults report using AI tools such as ChatGPT, Copilot or Gemini, up from 31% in 2024, per Ofcom. [2]
• Generative AI platforms drew an average of 9.5 billion monthly web visits worldwide between June 2025 and May 2026, up 70% year on year, on Similarweb's measurement. [3]
• ChatGPT held roughly 53% of worldwide AI web traffic in that window, down from around 76% the year before, with Gemini at 27-28% and Claude near 9%. [3][4]
• Google's AI Overviews appeared in 43% of US searches by July 2026, up from about 15% a year earlier, on Similarweb data reported by TechCrunch. [5]
• 68.01% of US Google searches ended without a click in January to April 2026, versus 60.45% in 2024, per SparkToro's analysis of Similarweb clickstream data. [6][7]
• Google AI Mode passed 1 billion monthly active users, and the Gemini app reached 950 million, Alphabet told investors in its Q2 2026 call. [8]
• ChatGPT reached 900 million weekly active users in February 2026, up from 800 million in October 2025, according to OpenAI. [9]
• Google still held 91.31% of worldwide search engine referrals in July 2026 on Statcounter's panel. [10]
• AI-referred retail visitors converted at a 54% higher rate than non-AI traffic in May 2026, reversing a large deficit from late 2024, per Adobe Digital Insights. [11]
• 65% of enterprise marketing leaders are putting at least a quarter of their 2026 budget into AI search optimisation, in a survey of 300 leaders published by Branch. [12]
• 45% of marketing leaders say they cannot accurately measure brand visibility in AI-generated answers, per Semrush's 2026 AI Visibility Index. [13]

How big is AI search in 2026?

Big enough to matter, and still much smaller than traditional search. Roughly half of adults in both the US and UK now use AI chatbots at all, and generative AI platforms collectively pull around 9.5 billion web visits a month worldwide.

The Pew Research Center surveyed 5,119 US adults in February 2026 and found 49% using AI chatbots, with 42% of users citing information search as a use case. [1] Ofcom put UK adult usage at 54% from a 7,533-person sample fielded in late 2025. [2]

Similarweb measured 655 million monthly unique visitors across generative AI platforms between June 2025 and May 2026, up 57% year on year. [3]

Here is the trap. "Half of adults use chatbots" is a reach number, not a volume number. Reach counts anyone who has used the tool; volume counts how many information-seeking queries actually get routed there instead of to a search box.

On volume, traditional search still dwarfs AI. Google alone held 91.31% of search engine referrals in July 2026 on Statcounter's panel, and Alphabet reported 17% year-on-year growth in Search and Other revenue in Q2 2026. [10][8] A channel does not grow revenue 17% while being replaced.

How fast is AI search actually growing?

Fast, from a small base, and the growth is concentrated in a handful of products. Platform-level visit growth ran at 70% year on year to May 2026, while individual apps posted triple-digit gains.

Comscore recorded 54.3 million mobile AI assistant visitors in December 2025, up 107% year on year, against 83.0 million on desktop, up 18%. [14] The mobile figure is the one to watch: it is where habitual, low-friction querying happens.

Source: Similarweb, reported July 2026. [4]

Percentage growth is flattering to small platforms. Claude growing 349% and ChatGPT growing 87% does not mean Claude is catching ChatGPT, because ChatGPT's 87% is applied to a far larger base.

The two declines are more interesting than the gains. Microsoft 365 Copilot and DeepSeek both shrank, which suggests bundling a model into an existing product is not the same as winning a query habit.

Which platforms hold what share of AI search?

ChatGPT still leads on web traffic, but its lead has roughly halved in two years. Gemini is the clear number two and the fastest riser among the large players.

Source: Similarweb, June 2025 to May 2026 measurement window. [3][4]

On self-reported usage the picture differs, because people use several tools. Pew found 44% of US adults use ChatGPT, 24% Gemini, 17% Copilot, 14% Meta AI, 8% Grok and 6% Claude. [1] Those add to well over 100% for a reason: multi-homing is normal.

Query volume is harder still. Perplexity disclosed 780 million monthly queries in May 2025 and claimed 20% month-on-month growth, but has not published a comparable verified figure since. [15]

For marketers the practical read is this. If you optimise only for ChatGPT, you are covering roughly half the traffic and a smaller share of the commercially valuable queries, because Gemini traffic increasingly arrives through Google surfaces where your existing SEO work still counts. This overlap is exactly what the generative engine optimization data digs into in more detail.

How much of Google search is now AI-generated?

Close to half of US searches now carry an AI Overview, and Google's dedicated AI Mode has passed a billion monthly users while still being a small fraction of total query volume.

Similarweb data reported by TechCrunch in July 2026 put AI Overviews on 43% of US searches, up from around 15% a year earlier. [5] AI Mode visits grew from 126 million in June 2025 to 279 million in May 2026 on the same measurement. [5]

Yet SparkToro found AI Mode accounted for just 0.34% of Google searches in its January to April 2026 US panel. [6]

Both things are true. A billion people have tried AI Mode; almost nobody uses it as their default search interface yet.

That gap between announced users and observed query share is the single most misused statistic in this space. "1 billion monthly active users" is a Google-reported engagement metric from the Alphabet Q2 2026 earnings call; 0.34% is an independent panel's measure of actual query routing. [8][6] Citing the first without the second overstates the shift by two orders of magnitude.

How many searches now end without a click?

More than two in three, in the US. The zero-click search statistics are the mechanism through which AI search damages publishers, not referral cannibalisation.

SparkToro measured 68.01% of US Google searches ending without a click between January and April 2026, against 60.45% in 2024, a rise of 7.56 percentage points. [6] Search Engine Land noted that click-generating searches fell 9.51 points across the same comparison. [7]

The click-through picture is not uniformly bleak. Seer Interactive found organic CTR on informational queries without an AI Overview rose from 2.93% in 2025 to 3.97% in early 2026, and reported an 85% rebound in AIO organic CTR over the first two months of 2026. [16]

Read that carefully. Fewer searches send clicks, but the clicks that do get sent may be arriving on pages that satisfy a more specific need, because the easy questions were absorbed upstream.

That is a volume problem dressed as a quality opportunity. Total sessions fall; value per session rises. Whether that trade is good depends entirely on whether your business monetises impressions or outcomes.

Is AI search traffic actually worth anything?

Per visit, yes, and by a widening margin. The AI referral traffic statistics show AI-referred visitors now outperform other channels on almost every engagement measure Adobe tracks, having badly underperformed them eighteen months earlier.

Adobe Digital Insights reported that in May 2026 AI-referred retail visitors converted at a 54% higher rate than non-AI traffic, were 36% less likely to bounce, spent 53% more time on site, viewed 23% more pages and generated 53% more revenue per visit. [11] In October 2024 non-AI visits had outperformed AI visits by 128% on conversion. [11]

Sector traffic growth to May 2026 ran at +194% for travel, +138% for retail and +105% for financial services. [11]

The volume side remains modest. Ahrefs, studying 3,000 websites in February 2025, found AI chatbots supplied 0.17% of visitors, with ChatGPT, Perplexity and Gemini accounting for 98% of that sliver. [17] That baseline is old now, but it establishes how low the starting point was.

The asymmetry is stark on the crawling side. Cloudflare measured crawl-to-refer ratios in August 2025 of 50,000:1 for Anthropic, 887:1 for OpenAI and 118:1 for Perplexity, with roughly 80% of AI crawling done for training. [18]

So: AI platforms take a great deal of content and return very few visits, but the visits they do return are unusually good. For a publisher funded by advertising that is a terrible bargain. For a B2B firm selling a considered purchase, it may be an excellent one.

What are marketers actually doing about AI search?

Budgeting for it, mostly, and struggling to measure it. Enterprise adoption of AI search optimisation is close to universal in stated intent, and much shakier in execution.

A survey of 300 enterprise marketing, growth and digital leaders published in April 2026 and reported by Business of Apps found 98% either already optimising for AI search or planning to within a year, 65% allocating at least 25% of 2026 budget to it and 28% allocating more than half. [12]

The same survey found 26% cannot track the journey from AI discovery to conversion, and 24% say their analytics tools are not ready for AI attribution. [12] Semrush, analysing 126 million US AI search prompts from January to April 2026, put the measurement gap at 45% of marketing leaders. [13]

Visibility is also brutally concentrated. Semrush found the top three brands took 82.9% of AI visibility in news and media and 76.9% in consumer electronics, though finance was more open at 41.4%. [13] Only 36 global brands held top-100 visibility across every platform every month. [13]

Bain & Company, surveying 1,500 US consumers in September 2025, found 44% of online buyers now mostly start in an LLM or split their search between AI tools and traditional engines. [19] Its analysis of roughly 500 million citations found 89% of unbranded prompts are answered using third-party sources. [19]

That last figure is the most actionable number in this article. If nearly nine in ten unbranded questions get answered from someone else's page, your owned content is not the lever. Third-party presence is, which is why the digital PR statistics matter so much for AI visibility.

How does the UK compare with the US?

Broadly similar on adoption, slightly ahead on exposure to AI summaries, and noticeably more sceptical about trusting the output.

Sources: Pew Research Center [1]; Ofcom [2].

The trust data is where the UK diverges most. Ofcom found 57% of aware adults would trust an AI-generated news story less than a human-written one, against 27% who would trust it equally and 7% who would trust it more. [2] Pew separately found 71% of Americans expect AI to make personal information less secure. [1]

High usage alongside low trust is not a contradiction. It describes a tool people find convenient and do not yet believe. For anyone serving a local market, the same convenience-versus-trust tension shows up in the local SEO statistics.

How to read these numbers

Panel measurement and server-side measurement do not agree, and neither is wrong. Similarweb, Statcounter and Comscore infer behaviour from panels and clickstream data, which under-represents in-app and mobile search app activity. Adobe measures actual visits on client sites, which is precise but limited to Adobe's customer base and skewed towards large retail and travel brands.

Self-reported survey figures are the softest. Pew and Ofcom are asking people to recall behaviour, and people overstate novel technology use.

Platform-announced numbers are marketing. "Monthly active users" for AI Mode and Gemini come from Alphabet's investor communications and are not independently auditable, which is why SparkToro's 0.34% query-share figure sits so awkwardly beside the billion-user claim. [8][6]

Several figures here are forecasts, not observations. eMarketer projects 31.3% of the US population will use generative AI search in 2026. [20] Gartner predicted in February 2024 that search engine volume would fall 25% by 2026 [21], and that has not happened on any public measurement: Google's share sat at 91.31% in July 2026 and its search revenue grew 17% year on year. [10][8] Treat confident directional predictions accordingly.

Finally, watch measurement windows. Similarweb's 53% ChatGPT share is a twelve-month average to May 2026; its June 2026 monthly reading was 52.7%. Averages and spot readings get quoted interchangeably and should not be.

What this means going into 2027

AI search is additive before it is substitutive. Google's query volume and revenue both grew while AI usage exploded. Plan for a larger total pie with a differently distributed slice, not a zero-sum transfer.

The click is the wrong unit. With 68% of searches ending without one and AI-referred visits converting 54% better in retail, session counts are becoming a poor proxy for commercial value. Attribution needs rebuilding around assisted and unattributed discovery.

Third-party presence beats owned content for AI visibility. If 89% of unbranded prompts resolve to third-party sources, digital PR, review platforms and comparison sites do more for AI visibility than another landing page.

Measurement will be the differentiator, not tactics. With 45% of marketing leaders unable to measure AI visibility and 26% unable to trace AI-sourced conversions, the competitive edge in 2027 goes to whoever can prove the channel works.

Expect concentration to persist. Top-three brands already hold over 80% of AI visibility in some categories. Challenger brands should target the fragmented ones, such as finance at 41.4%, rather than fight where the answers have already calcified.

If you want a second pair of eyes on where your brand currently stands across the AI engines, and a concrete plan to earn visibility where it counts, you can book a call with me and we will look at your numbers together.