AI Overviews UK Statistics: The 2026 Numbers

Your organic traffic is down, someone has forwarded you a percentage to explain it, and now you have to decide whether that percentage should change how you spend money. Most of the AI Overviews UK statistics doing the rounds in business groups and agency pitch decks arrived with no source, no date and no method attached. That makes them useless for a decision and actively risky in a conversation with a business partner or a board.

This post is the numbers, and only the numbers. If you want the explanation of what AI Overviews are and how Google assembles them, our explainer on Google AI Overviews for Suffolk businesses does that job properly. What you get here is a dated table of what has genuinely been measured in the UK, who measured it, what each figure covers and what it quietly leaves out, and then the part that actually changes your decisions: how to work out your own number instead of borrowing somebody else's average.

Last checked 6 September 2026. Every figure below carries the month it was measured, because a search statistic with no date is a rumour.

Key takeaways

• Ofcom's Online Nation report, published December 2025, found around 30% of UK searches display an AI Overview and 53% of UK adults say they see AI summaries regularly.
• SparkToro and Similarweb measured 69.5% of UK Google searches ending without a click between January and April 2026, the highest zero-click rate of the six countries in the study.
• Ahrefs found the top organic result keeps only 42 of every 100 clicks it used to get when an AI Overview appears, a 58% fall, measured in February 2026.
• Seer Interactive measured brands cited inside an AI Overview earning roughly 35% more organic clicks in September 2025, widening to about 120% more clicks per impression by April 2026.
• The widely quoted claim that AI Overviews appear on 41% of UK searches has no named study behind it and should not be used to justify spend.

Where does the 41% figure actually come from?

It gets repeated constantly, usually as a jump from around 14% in late 2025. We went looking for the study. There is not one. No named research organisation, no sample size, no measurement window, no methodology note. It appears in roundup posts citing other roundup posts, which is how a guess becomes a fact.

What does exist is a set of real measurements that disagree with each other, and understanding why they disagree is more useful than any single number. Ahrefs put AI Overviews on 9.46% of desktop keywords in May 2025. Semrush tracked a range across 2025 that settled near 15.7%. Conductor reported 25.11% in early 2026. BrightEdge reported roughly 48% of its tracked queries in February 2026. Whitespark found 68% of local business queries carried one. Those are not contradictions. They are five different keyword sets.

The distinction that matters: a percentage of tracked keywords is not a percentage of searches. A keyword set is a sample a vendor chose, usually weighted towards the informational terms their customers buy tools to rank for, and those are exactly where AI Overviews fire most. A percentage of searches is what people actually typed, weighted by real volume, most of which is brand names, navigational queries and short local intent. Quote the first as though it were the second and you overstate your exposure by a factor of three.

What has actually been measured in the UK

Four figures survive scrutiny. Three come from Ofcom, a regulator publishing documented methodology rather than a vendor publishing a lead magnet, and one comes from clickstream panel data.

The zero-click figure is the one that should worry you. The UK topped that study: Germany 62.1%, Italy 63.4%, Canada 63.8%, France 65.3%, the United States 68.0%. British searchers click out to the open web less than any of them. Per 1,000 UK searches, only 232 clicks reached a site Google does not own.

Why Ofcom is the number to put in front of a partner

If you are arguing for a budget change, cite the regulator, not a vendor blog. Ofcom publishes its sampling and its caveats, and nobody in the room can dismiss it as marketing for a subscription product. The trade-off is that regulator data is slower, so the December 2025 figure is nine months old and the real rate has almost certainly risen since. Say that out loud rather than pretending the number is live.

How far have click-through rates actually fallen?

The top organic result

Ahrefs published the sharpest measurement on 6 February 2026, and it is worth quoting the shape of it rather than the headline: for every 100 clicks that would once have gone to the first-ranking page, 42 remain. Google keeps the other 58. Second place lost 50.8% and third place lost 46.4% over the same comparison. The earlier Ahrefs study, comparing March 2024 with March 2025 across 300,000 keywords, put the drop at 34.5%. So the decline is not just real, it accelerated across 2025.

Pew Research, working from a US browsing panel in July 2025, found users clicked a link on about 8% of visits to a results page carrying an AI summary, against 15% without one. Only 1% clicked a link inside the Overview. That last number is uncomfortable for anyone selling citation chasing as a traffic strategy.

What being cited is worth

Seer Interactive has run the most consistent measurement of the upside. In September 2025 it found brands cited inside an AI Overview earned roughly 35% more organic clicks, and about 91% more paid clicks, than uncited brands on the same query. By April 2026 that gap had widened to around 120% more organic clicks per impression. Seer also recorded click-through on Overview queries recovering from 1.3% in December 2025 to 2.4% in February 2026, so the picture is still moving.

One caveat that never travels with the statistic: these are brand-level comparisons on queries with enough impression volume to measure. A village butchery near Bury St Edmunds getting named in an Overview for one query a fortnight is not going to see a 120% anything. The direction is right. The magnitude is not yours.

Why the headline trigger rate overstates your exposure

BrightEdge tracked AI Overview coverage by sector between February 2025 and February 2026 and found healthcare queries at 88% coverage while ecommerce and transactional queries sat at roughly 4%, down from 29% a year earlier. Google appears to have deliberately pulled Overviews back from queries where somebody is about to buy something. That is the single most useful fact in this post for a small business.

• Informational and advice queries. Highest exposure by a distance. If your traffic came from how-to content, it has gone and it is not coming back.
• Local business queries. Whitespark found 68% carry an Overview against 39% carrying a traditional local pack, but these Overviews usually name businesses, so the question is whether you are one of them.
• Transactional and product queries. Lowest exposure, and falling. The searcher who types a service plus a town and is ready to ring somebody is still mostly seeing links.
• Branded queries. Around a 20% trigger rate against roughly 80% for non-branded, which is one more argument for building a name people search for by itself.

Run those four against your own enquiry mix. A Newmarket clinic whose enquiries come from a treatment plus the town is barely exposed. A firm whose traffic came from a library of explainer articles has already lost most of it. Same month, opposite problems, and one national percentage tells neither of them anything.

AI Overviews UK statistics cannot tell you what happened to your traffic

National averages hide everything that matters at your scale. Your exposure depends on the mix of query types that actually produce enquiries, on your sector, on how much of your traffic is people typing your name, and on whether you sell something Google has decided to summarise. None of that is in a 30% or a 69.5%.

There is also a timing trap. Rankings moved in the same window for reasons unrelated to AI. If your traffic fell in late May or early June 2026 you cannot attribute it to Overviews without first ruling out the ranking reshuffle described in our post on recovering from a Google core update. Attributing a core update loss to AI Overviews leads to the wrong fix, and the wrong fix costs a quarter.

How to work out your own number in Search Console

Google rolled its generative AI performance reports out worldwide on 31 August 2026, so this is now available to every UK site rather than a handful of early markets. Read the limitation before you read the data: the report gives impressions, pages, countries, devices and dates, and it does not give click data at all. It tells you where you are appearing, not what appearing is earning you. Our guide to measuring AI search traffic in Search Console goes through the interface in detail.

1. Set a 16-month comparison in the Performance report so you are looking at year on year, not month on month noise.
2. Work at query level, not sitewide. Sitewide click-through rate averages away the exact signal you are hunting for.
3. Open the generative AI performance report and list the pages picking up AI impressions.
4. Cross-reference the two. Pages with rising AI impressions and falling click-through rate are your genuinely exposed pages. That list is usually shorter than owners expect.
5. Hand-check five money queries from a UK phone, logged out, and record whether an Overview appears and which businesses it names.
6. Repeat quarterly and write the date on it. A number without a date is the problem this whole post is about.

Step five is not busywork. With no click data in the report, the manual check is the only way to see whether the Overview on your core query names you, a competitor, or Checkatrade.

What these numbers justify spending money on

They justify finding out which pages are exposed and fixing those. They justify writing pages that answer the questions people ring you about, with prices and service areas in plain words, because that is what gets quoted. They justify research that steers away from query types Google now answers itself, which is the point of content planning before anything gets written.

They do not justify a tool subscription bought on the strength of a percentage nobody can source. They do not justify chasing citations on informational queries that never produced an enquiry. And they do not justify blocking Google's AI features out of principle: Google confirmed in August 2026 that the opt-out control is not used as a ranking signal outside the AI surfaces, but opting out also means no impressions and no traffic from them, which is a trade with no upside for a local business that wants to be named.

For a firm covering Cambridgeshire from Cambridge out to the Fens, the query mix differs town by town, and the account average tells you nothing about which town is losing ground.

Frequently asked questions

What percentage of UK searches have AI Overviews?

Around 30%, according to Ofcom's Online Nation report published in December 2025, which is the most credible UK-specific measurement available. Vendor studies range from 9% to 68% because each tracks a different keyword sample. The true figure has probably risen since December 2025, and no regulator has published an updated UK count as of September 2026.

Is the 41% AI Overviews figure wrong?

It is unsourced rather than provably wrong. No named research organisation, sample size or measurement window has ever been attached to it in the UK marketing content that repeats it. Do not build a budget case on it. Ofcom's 30% and the vendor ranges are defensible in a room where somebody might ask where the number came from.

How much traffic have AI Overviews cost small businesses?

For the top organic result on an affected query, Ahrefs measured a 58% fall in click-through rate in February 2026, up from 34.5% in its 2025 study. Whether that applies to you depends entirely on query type: informational content has been hit hardest, transactional and local service queries far less, and branded queries barely at all.

What is the UK zero-click search rate in 2026?

69.5% of UK Google searches ended without a click on data collected between January and April 2026, published by SparkToro using Similarweb panel data in June 2026. The UK ranked highest of the six countries measured, ahead of the United States at 68.0% and Germany at 62.1%. Only 232 clicks per 1,000 UK searches reached a non-Google site.

Does being cited in an AI Overview bring clicks?

Yes, measurably, but less than the headline suggests. Seer Interactive found cited brands earning around 35% more organic clicks in September 2025 and roughly 120% more per impression by April 2026. Pew Research, though, found only 1% of users clicked a link inside the Overview itself, so most of that benefit is indirect rather than a direct click.

How do I measure the impact on my own site?

Use the generative AI performance report in Search Console, available worldwide since 31 August 2026, alongside a year on year query-level click-through comparison. The AI report shows impressions, pages, countries, devices and dates but no click data, so pair it with hand-checking five of your commercial queries from a UK phone each quarter.

What we would do with these numbers on a Monday morning

Open Search Console, pull the twenty queries that actually produce enquiries, and check the click-through rate on each against the same period last year. Most owners we do this with in Bury St Edmunds and Ipswich find two or three pages have genuinely lost ground and the rest are flat. That is a very different problem to the one the 41% headline implies, and it costs a great deal less to fix.

Then decide what the exposed pages were ever for. If a page existed to catch informational traffic Google now answers itself, rewriting it will not bring that traffic back, and the honest move is to repoint the effort at queries where somebody is ready to buy. That is a content decision, and it belongs in the ongoing search programme inside every plan rather than a separate AI product.

Then put a date on your own numbers and recheck in January. A statistics page nobody has revisited in a year is worth less than no page at all, which is why this one carries a checked date at the top.

If you would like us to run the query-level comparison on your own Search Console data and tell you which pages are genuinely exposed, ask us for a traffic review.