AI Search Statistics for 2027: AI Overviews, ChatGPT and What Happens to Clicks

Over one recent 90-day stretch, a blog post of ours picked up 39,208 impressions in Google Search Console but only 13 clicks. The post sat at position 2.9, which sounds like a dream, yet almost nobody came to the page. Digging in showed the cause had nothing to do with AI (we explain it in our Search Console data study), but the episode taught us something that applies directly to AI search: impressions, rankings and clicks no longer move together, so you need to know which of them you're reading before you decide anything.

AI answers are a normal part of search now. Google shows AI Overviews to billions of people, ChatGPT has hundreds of millions of weekly users, and independent research finds that people click ordinary results less when an AI summary sits on the page. This page gathers the AI search figures that come from the companies themselves or from independent research bodies, and it explains what they mean for a business that relies on search. We left out vendor surveys and predictions such as "AI will replace Google by 2028".

AI search statistics at a glance

How many people use AI search?

Billions. Google said in July 2025 that AI Overviews had over 2 billion monthly users in more than 200 countries and territories and 40 languages, and that AI Mode had over 100 million monthly users in the U.S. and India. OpenAI said ChatGPT reached 800 million weekly users in October 2025.

In Alphabet's Q2 2025 results, Google's chief executive put it directly: "AI Overviews now has over 2 billion monthly users across more than 200 countries and territories and 40 languages." AI Mode, the chat-style search experience, "already has over 100 million monthly active users in the U.S. and India."

Outside Google, OpenAI announced at its October 2025 developer event that ChatGPT had reached 800 million weekly active users. In the U.S., the Pew Research Center found that 34% of adults had used ChatGPT by early 2025, about double the share in 2023. Among adults under 30 it was 58%, and 28% of employed adults said they use it for work.

The UK numbers show how fast this moved. Ofcom reported that ChatGPT had 1.8 billion UK visits in the first eight months of 2025, up from 368 million in the same period of 2024. That's almost five times as many in a single year.

What happens to clicks when an AI summary appears?

Clicks drop when an AI summary appears. In Pew Research Center's tracking of real browsing, visits with an AI summary led to a click on a traditional result 8% of the time, versus 15% for visits without one. Only 1% of visits included a click on a link inside the summary, and people more often ended the session there.

Pew Research Center's July 2025 study is the most careful independent measurement we have so far. Pew drew on the real browsing of 900 U.S. adults in March 2025, which covered 68,879 unique Google searches, and a few results are worth a look:

• 18% of searches ended up with an AI summary. Longer searches phrased as questions triggered one more often, while searches of one or two words produced one just 8% of the time.
• With a summary on the page, a traditional result got clicked on 8% of visits; with none, the click rate was 15%.
• A link inside the AI summary was clicked on just 1% of visits.
• Sessions ended on the page for 26% of visits with a summary, compared with 16% when there wasn't one.

Google disagreed with the study, arguing that its method and its set of queries didn't represent search traffic. That's worth knowing, but it's still the best independent measurement that's public, and it fits what many site owners describe: impressions hold steady while clicks fall, and informational questions lose the most.

Picture a typical service business. A homeowner types in "why is my boiler losing pressure". A few years back they'd have opened two or three articles, but now the AI summary probably answers them right on the results page. A simple problem gets fixed on the spot, and nobody receives a visit. A harder one sends them back to search, this time looking for a business, and on that second search your profile, your reviews and your brand name decide which company they phone.

What does Google say you need to do?

Nothing special. According to Google's documentation, showing up in AI Mode or AI Overviews takes no additional requirements and no special schema markup. Your page only needs to be indexed and able to qualify for a snippet, and Search Console reports clicks from AI features inside its regular Web search totals.

The page Google publishes on AI features and your website is brief and blunt: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." Your page needs to be indexed and able to appear in Google Search with a snippet, and, in Google's words, "there's also no special schema.org structured data that you need to add."

Google also covers measurement. Clicks from AI features "are included in the overall search traffic in Search Console", under the Web search type, so you can't pull them out separately there. To see which AI tools send you visitors directly, our guide to tracking AI traffic in GA4 walks through the setup.

That leaves plenty of room for snake oil. If someone sells "AI Overview optimisation" as its own technical service, keep in mind that Google says nothing special is needed on that front. What works is what has always worked: content written clearly for people that's accurate and easy to follow. Google's own guidance on helpful content asks whether content is "created primarily for people, and not to manipulate search engine rankings."

What kind of pages get cited?

Pages that state the answer first, name a source for every number, use tables and lists, and show who wrote them and why they know the topic tend to get cited. We can't see inside any AI system, so this is only our own observation, not a set of ranking factors.

1. The answer comes first. Give one or two plain sentences that settle the question first, with background after. It's the reason each section on this page begins with a brief, direct answer.
2. Numbers carry their source. When a figure comes with a publisher's name and a link, readers can trust it more easily, and it's simpler to quote than a claim with no source.
3. Tables and lists. Comparisons set out in a clear structure are simple to pull into an answer, and readers can skim them fast.
4. A clear author and business behind the page. Google's guidance on helpful content raises the question of whether it is "self-evident to your visitors who authored your content."
5. Something original. Use your own data, process and pricing, because when your page repeats what everyone else's says, an AI summary has no reason to choose it.

If you want the wider strategy, SEO vs AEO vs GEO is the one to read.

Before and after: an answer-first rewrite

Here's what "put the answer first" looks like on a real kind of page. Take a typical opening paragraph from a plumbing company's boiler service page:

Before: "At ABC Plumbing we have been serving the community for many years and pride ourselves on our friendly, reliable team. We understand that your boiler is one of the most important appliances in your home, which is why we offer a full range of services to keep it running smoothly all year round."

It's pleasant, but it doesn't answer anything. Someone who searched "what happens during a boiler service" learns nothing from it, and neither does an AI system looking for a quotable answer.

After: "A boiler service with us takes about an hour. We check the burner, the flue, the pressure and the safety devices, clean what needs cleaning, and give you a written record you can keep for your warranty and insurance. If we find a fault, we explain it and quote before doing any extra work."

The second version answers the question in its first sentence, adds a practical detail and only then talks about the business. If an AI summary quotes either paragraph, it'll be the second. And if a person reads it, they're more likely to trust a company that answers straight. (The business and wording are an illustration, not a real client.)

You don't need to rewrite your whole site at once. Start with the five pages that bring the most search impressions in Search Console, rewrite their first paragraph so it answers the main question directly, and compare clicks and enquiries over the following two months.

How AI search shows up in your data

Google folds clicks from AI features into its regular search totals, so Search Console has no separate line called "AI Overviews". You may spot a pattern instead: impressions steady or climbing, clicks dropping, and average position staying put. On its own that pattern doesn't prove AI summaries are responsible, because seasonality and competitors can produce it as well, but when it appears mostly on question-style searches, AI answers are likely part of the cause.

Referrals that come straight from AI are easier to find. GA4 records visits from chat assistants as referral traffic from their domains, and you can gather them into a custom channel to see them at a glance. There usually aren't many, but they're often well qualified, since these visitors have asked detailed questions before they click. Our guide to GA4 vs Search Console shows why the two tools won't ever report matching figures here.

Figures you'll see elsewhere that we left out

Pages of AI search statistics are often padded with forecasts ("AI search will overtake traditional search by 2028"), claims about how much more AI visitors are worth, and click studies run by tool vendors on their own panels. A few may well prove right. That's why they aren't here: either they're predictions or we couldn't verify the method behind them. Everything above was measured, published by the group that measured it, and each comes with a link you can follow to verify it.

What should a business do about AI search?

Carry on with SEO, since AI features draw on pages that are indexed. Then adjust what you measure by tracking enquiries and brand searches as well as clicks. Rewrite your main pages and put the answer at the top, cite a source for each number, and build up your Business Profile and reviews for the point when people look up your business by name.

1. Measure leads, not just clicks. If clicks fall but calls and form fills hold steady, AI summaries are taking the quick-question visits you weren't converting anyway. Our GA4 key events guide shows how to track real enquiries.
2. Watch brand searches. Searches for your business name are the ones AI can't answer for you. Growing them is the best insurance there is.
3. Put answers first on your key pages. Open each service page with two sentences answering the main question, and put the detail after them.
4. Add original material. Prices, timelines, before-and-after results and your own data.
5. Check how your pages are built. Our free website analysis will show you gaps in technical setup and in content.
6. Review monthly, not daily. AI features change often, so check your numbers against the previous month and against the same month a year earlier.

Sources and method

The figures here come from Search Central documentation and Alphabet's Q2 2025 results (both Google), plus OpenAI's public announcement, Ofcom and Pew Research Center, and we checked each one on its source page. Our own Search Console numbers are from greatinflux.com. Third-party traffic estimates and forecasts are left out.