More than half of UK adults, 54%, now use AI tools such as ChatGPT, Copilot or Gemini, up from 31% in 2024, according to Ofcom's own 2025 research. ChatGPT alone recorded 1.8 billion UK web visits in the first eight months of 2025, up from 368 million in the same period a year earlier, and roughly 30% of Google searches in the UK now trigger an AI-generated summary. For any UK business thinking about content strategy, this is no longer a forward-looking trend to plan for. It is the present-day baseline.
Ofcom, the UK's communications regulator, publishes two separate research instruments that both touch on AI adoption, and conflating them produces misleading numbers. This article works from Ofcom's own published figures, verified directly against ofcom.org.uk, to set out what is actually known about how UK adults use AI search today.
The Adoption Curve: From 23% to 54% in Two Years
UK consumer AI adoption rose from 23% of adults in 2023 to 31% in 2024, then reached 54% in late 2025, according to Ofcom's Adults' Media Use and Attitudes Tracker, a survey of 7,533 UK adults aged 16 and over. A separate Ofcom instrument, Online Nation, independently reports a comparable figure of up to 56% of online adults using AI tools, a rise of roughly 14 percentage points in a single year. Because these are two different survey instruments measuring a similar but not identical question, Ofcom's own guidance treats 54% to 56% as a band rather than a single precise figure, and this article does the same.
The generational concentration behind that average is stark. Ofcom found 79% of UK adults aged 16 to 24 and 74% of those aged 25 to 34 currently use AI tools, against only 18% of adults aged 65 and over. A business whose customers skew older cannot assume the 54% national average applies evenly to its own audience, and one whose customers skew younger should assume AI-mediated discovery is closer to the default than the exception already.
ChatGPT's Growth, and What Ofcom Did Not Disclose
ChatGPT's UK traffic growth is the best-documented single data point in this entire research area. Ofcom's Online Nation 2025 report recorded 1.8 billion UK visits in the first eight months of 2025, up from 368 million in the same period of 2024, with 252 million UK visits in August 2025 alone, a 156% year-on-year increase. Rival platforms also grew over the same period: Gemini's UK visits rose 146%, Claude's rose 138%, and Perplexity's rose 100% in the year to August 2025, though from much smaller starting bases.
Here is the detail that matters for anyone quoting these figures: Ofcom explicitly did not disclose the absolute visit or user counts behind those percentage increases for Gemini, Claude or Perplexity. Any published figure claiming a precise UK monthly-active-user number for those three platforms is coming from a different source, typically a commercial analytics vendor, and should be labelled as such rather than attributed to Ofcom. Separately, UKOM's Ipsos iris Online Audience Measurement Service, cited via IAB UK, put ChatGPT's UK monthly audience at 16 million in September 2025, a figure from a different, named measurement methodology than Ofcom's visit-count data, and the two should not be merged into one number either.
The Zero-Click Shift: What 30% AI Overview Coverage Actually Means
Ofcom found that around 30% of keyword searches in the UK now deliver an AI-generated response, and 53% of adults say they see these summaries often, whether or not they were looking for them. This is a passive adoption pattern: most people encountering AI Overviews did not seek out an AI product, they searched normally and the AI layer appeared inside their existing search results.
The behavioural consequence is a documented decline in click-through to original content. Ofcom's own analysis, using Similarweb data as a proxy from the US market, found that six months after a user's first visit to ChatGPT, their Google search clicks reduce by 26%. Separately, Semrush's analysis found a roughly 40% reduction in click-through rates on organic search results when an AI Overview is present for that query. These are two different studies measuring related but distinct effects, one about a user's broader search behaviour after adopting ChatGPT, the other about click-through on a specific results page, and both point in the same direction without being interchangeable statistics.
| Metric | Figure | Source |
|---|---|---|
| UK adults using AI tools | 54% (up from 31% in 2024) | Ofcom Adults' Media Use Tracker 2025 |
| ChatGPT UK visits, first 8 months 2025 | 1.8 billion (vs 368 million in 2024) | Ofcom Online Nation 2025 |
| ChatGPT UK visits, August 2025 alone | 252 million (+156% YoY) | Ofcom Online Nation 2025 |
| Google searches showing AI summaries | ~30% | Ofcom Online Nation 2025 |
| Adults who say they see AI summaries often | 53% | Ofcom Online Nation 2025 |
| ChatGPT UK monthly audience, Sept 2025 | 16 million | UKOM Ipsos iris via IAB UK |
What UK Adults Actually Ask, and Which Sources AI Cites Back
UK AI search queries look different from a typical Google keyword search. Users ask longer, more conversational questions, often with a regulatory or institutional anchor built in, such as questions about HMRC rules for contractors, GDPR obligations for a specific business scenario, or evidence-based comparisons of professional service providers. This is a distinct pattern from the shorter, keyword-centric queries that still dominate in less digitally mature markets, and it has a direct implication for what gets cited back.
A distinctive feature of the UK AI search environment is the elevated role of nationally specific institutional authority. GOV.UK and NHS sources dominate citations for regulatory and health queries in Perplexity, while BBC, The Guardian, City A.M. and The Drum are cited disproportionately for UK-specific queries in ChatGPT Search. Content that references UK-specific regulations, government bodies and recognised industry standards is structurally favoured in these responses over generic international content that omits British context, a pattern independently confirmed across UK-focused GEO market commentary.
The Adoption Gap Behind the Headline Number
The 54% consumer figure sits alongside a separate and considerably lower business-adoption figure, and the two should never be quoted interchangeably. The UK's Department for Science, Innovation and Technology found that only 16% of UK businesses with five or more employees formally used at least one AI technology in 2025, using a strict definition of formal adoption. A separate ONS Business Insights and Conditions survey, using a broader definition, put the figure at 23%. Large firms with 250 or more employees report adoption closer to 44%, while small firms under 50 employees sit nearer 26%, according to ONS data analysed by the University of Cambridge's Bennett School of Public Policy.
That means a UK business's own customers may already be searching with AI daily, while the business itself, and potentially its competitors, have not yet formally adopted AI tools in their own operations. This asymmetry, not the headline 54% figure on its own, is the more commercially useful data point: it identifies a specific gap between demand-side behaviour and supply-side readiness that a business can act on now, before competitors close it.
How Adoption Varies by Sector and Business Size
The 16% to 44% range covered above hides real variation once broken down by sector and by company size, and two separate UK government-adjacent sources report that breakdown slightly differently, which is itself informative. DSIT-sourced sector analysis puts information and communication firms at the top of business AI adoption, 43% to 51% depending on the benchmark used, followed by financial services at 21% to 31% and professional services at 20% to 28%. Construction and transport lag considerably, at roughly 6% to 12% and 10% respectively, the widest sector gap in the UK adoption data.
A separate breakdown, published via a UK government impact assessment and distinct from the ONS BICS figures cited earlier, reports formal AI adoption at 36% for large businesses, 23% for mid-sized businesses and 14% for microbusinesses, alongside a further finding that 25% of UK businesses overall were using at least one surveyed AI technology, up from 16% in January 2025. This is a different survey, with different timing and methodology, from the ONS BICS/Cambridge Bennett School figures of 44% for large firms and 26% for small firms cited earlier in this article. Both are legitimate, both are UK government-adjacent, and neither should be quoted as "the" UK business adoption figure without naming which survey it came from.
The commercially useful reading of this is not to average the two large-firm figures into a false middle number, but to recognise that large UK companies consistently adopt AI formally at two to three times the rate of small and micro businesses, regardless of which specific survey is cited, and that Information and Communications firms are consistently the most AI-mature sector across every UK data source that measures it.
What This Means for a UK Content Strategy
The practical implication is not that traditional SEO stops mattering. AI engines still source most of their answers from indexed, crawlable web content, so a weak organic foundation limits how much visibility a brand can gain inside an AI-generated summary regardless of how well the content is structured for citation. What changes is the destination the content needs to be built for: getting named accurately inside the answer itself, not just ranking for the click that increasingly does not happen.
Ofcom's own November 2025 discussion paper, "Era of Answer Engines," documents early evidence that publishers are already seeing reduced traffic as AI summaries divert users away from clicking through, a tension the regulator is actively monitoring rather than treating as settled. For a UK business, that makes the case for structuring content so that AI engines can extract and cite it accurately even when the user never visits the site directly, the exact discipline covered in our companion piece on how to evaluate genuine UK GEO capability.
We treat this shift as the starting premise for our own GEO service, and Tessar Napitupulu's Cited or Silent opens with exactly this argument: that AI-generated answers are no longer a future risk to plan for but a present-day distribution channel that already determines whether a brand is found before a customer ever searches for it by name.
Frequently Asked Questions
Is the 54% AI adoption figure the same as the percentage of people using AI to search?
No, and this is the single most common conflation in this data. The 54% figure measures general AI tool usage, including tasks like writing help or image generation. AI-search-specific behaviour, such as ChatGPT's 1.8 billion UK visits, is a separate, more specific measurement and should be cited independently.
Why do different sources report different numbers for ChatGPT's UK user base?
Because they measure different things. Ofcom's Online Nation report measures web visits using Similarweb-style data, which counts sessions, not unique users. UKOM's Ipsos iris service measures monthly audience reach using panel-based methodology. Both are legitimate, named sources, but they answer different questions and should never be merged into a single figure.
Does a 30% AI Overview trigger rate mean 30% of my potential customers will not see my website?
Not directly. The 30% figure is an average across all UK Google searches, and AI Overview trigger rates vary significantly by query type, with informational and how-to queries affected more than commercial or navigational ones. The relevant question for a specific business is what share of its own target queries currently trigger an AI Overview, which requires testing rather than applying the national average.
Should a UK business stop investing in traditional SEO given this shift?
No. AI engines still draw most of their answers from indexed, crawlable web content, so technical SEO remains the foundation that AI visibility is built on top of, not a separate or competing investment.
Why do UK AI Overview citations favour GOV.UK, NHS and BBC over other sources?
UK-focused AI search research consistently finds that nationally specific institutional sources are treated as higher-authority for regulatory, health and current-affairs queries. Content that references the same institutions accurately tends to perform better in UK-targeted AI citations than generic international content.
How can a UK business tell if its customers have already moved to AI-mediated search?
The most direct method is testing a representative set of the business's own buyer-intent queries across ChatGPT, Perplexity and Google AI Overviews and observing whether the brand, or its competitors, appear in the response. National adoption averages are a starting signal, not a substitute for testing a business's own actual query set.
Why do large-firm adoption figures range from 36% to 44% depending on the source?
Because they come from different surveys measuring different things. The 44% figure is from ONS Business Insights and Conditions Survey data analysed by the University of Cambridge's Bennett School. The 36% figure comes from a separate UK government impact assessment with its own survey methodology and timing. Neither is wrong; they are simply not directly comparable without naming the source.
Sources & References:
- Ofcom, "Online Nation 2025" report, published 10 December 2025, on ChatGPT UK visit volumes, AI Overview trigger rates and the post-ChatGPT click-through decline.
- Ofcom, "Passive social media use, AI companionship, and online side hustles" (Adults' Media Use and Attitudes Tracker), on the 54% AI adoption figure and age-group breakdown, fieldwork 29 September to 28 November 2025.
- Ofcom, "Ofcom's strategic approach to AI, 2026/27," on the 54% adoption figure cited against a 31% baseline in 2024.
- UKOM Ipsos iris Online Audience Measurement Service, via IAB UK, on ChatGPT's 16 million UK monthly audience figure for September 2025.
- Semrush, click-through-rate analysis on organic search results when Google AI Overviews are present.
- Ofcom, "Era of Answer Engines" discussion paper, November 2025, on publisher traffic concerns arising from AI summary adoption.
- UK Department for Science, Innovation and Technology (DSIT) sector-level adoption analysis and Helium42 benchmark data, on Information and Communications leading UK sector adoption at 43–51%, against Construction and Transport lagging at 6–12% and 10% respectively.
- UK government impact assessment (published via assets.publishing.service.gov.uk), on the 36%/23%/14% large/mid/microbusiness adoption breakdown and the 16%-to-25% adoption growth from January 2025.