Do AI Summaries Reduce Clicks, And What Does The Research Say?

Zero-click Existed Before AI

Zero-click search, where the user finds the answer on the results page and leaves without visiting any site, was a topic long before AI summaries. Weather, exchange rates, a shop’s closing time: the answers to these have sat on the page itself for years. SparkToro’s 2024 study, built on Datos panel data, calculated that 58.5% of Google searches in the US and 59.7% in the European Union ended without a click¹. The same study found that out of every 1,000 searches in the US, only 360 clicks went to the open web outside Google’s own properties¹.

So even before AI summaries arrived, the traffic reaching a business’s site was fed by only a share of searches. What has changed is that the answer kept on the results page has grown from a single line of information into an explanation several paragraphs long. In February 2024 Gartner predicted that traditional search engine volume would drop 25% by 2026 because of chatbots and other virtual agents². It was a forecast, not a measurement; still, it was enough to reach the meetings where marketing budgets are discussed.

What Pew Measured: What People Do When A Summary Appears

Some of the cleanest data in this area came from Pew Research Center. The researchers examined the real Google searches made in March 2025 by 900 US adults who had agreed to have their browsing tracked; 18% of those searches produced an AI summary³. On pages without a summary, 15% of visits ended with a click on a traditional result link, while on pages with one the figure fell to 8%³.

More striking is what happens to the source links inside the summary. Users clicked a link in the summary in only 1% of visits where one appeared³. On pages with a summary, 26% of visits ended the browsing session entirely; on pages without one, the figure was 16%³. People read the answer, put the phone down and get on with their day.

The study’s value lies in its method: actual behaviour, not a survey. Its limits are also clear; the sample is US users over a single month. We do not know whether users in Türkiye behave at the same rates, but we see no reason to expect the direction to differ.

Click-through Data Points The Same Way

Ahrefs compared 300,000 informational keywords with and without an AI summary. Between March 2024 and March 2025, the click-through rate of the top-ranking page on queries with a summary ended up about 34.5% below the level that the trend of similar queries would predict⁴. The authors state plainly that this is a correlation, not causation⁴.

Seer Interactive’s study rests on different data: Search Console records from 42 organisations and 3,119 informational search terms⁵. From June 2024 to September 2025, organic click-through on queries with a summary fell from 1.76% to 0.61%⁵. Queries without a summary declined too, from 2.74% to 1.62%⁵.

That second detail is easy to miss. The erosion of clicks cannot be explained by summaries alone; the results page is getting more crowded, people are asking longer questions, and some go to a chat assistant without touching a search engine at all. Summaries are the most visible part of this broader trend, but not the only part.

Why Google Tells A Different Story

Google’s position has differed from the start. Liz Reid, head of Search, wrote in May 2024 that links included in AI summaries get more clicks than the same page would have received as a traditional listing⁶. In August 2025 she said that total organic click volume from Google to websites had stayed roughly flat year on year, and that “quality” clicks, where the user does not quickly return, had risen slightly⁷. In the same post she attributed outside reports of sharp traffic declines to flawed methods and isolated examples⁷.

The difficulty in weighing these claims is that Google does not share figures. Total clicks are said to be stable, but we are not shown which data or which period is being compared. The independent studies, by contrast, speak from narrower samples but with their methods in the open.

Both sides could be right at the same time. If the total number of searches is growing, fewer clicks per search can still add up to a similar volume overall. Reid herself acknowledges that traffic is shifting between sites, with some gaining and others losing⁷. For a business the real question is not the overall picture but which side of that shift it stands on.

What Being Cited In The Summary Brings

The most interesting finding in Seer’s study is the gap between brands shown as a source inside the summary and those that are not. In the third quarter of 2025, brands cited in the summary had an organic click-through rate about 35% higher, and a paid click-through rate about 91% higher, than brands that were not cited⁵.

The researchers are careful here: they write that they cannot definitively prove citation drives clicks, and that strong, well-known brands may be more likely both to be cited and to be clicked⁵. So cause and effect have not been established. Still, this much can be said: a business not named in the summary is absent from the only text the user read during that search.

Picture a dental clinic. Someone who searches “how long does swelling last after an implant” may read the answer in the summary and close the page. They will not visit the site that day; but they may remember the clinic named in the summary when they look for an appointment a few weeks later. That kind of effect never lands on any line of a click report.

Why People Settle For The Summary

One reason behind the change in behaviour is simple: a short answer is enough for many questions. In a randomised experiment run by Xu, Feng and Chen in 2023, participants using a ChatGPT-like interface finished tasks faster than those using a Google-like one, and there was no significant difference in overall performance between the groups⁸. The ChatGPT group was more satisfied with the experience, but did worse on tasks that required fact-checking⁸.

This finding also explains where clicks are moving. Simple, definitional questions get resolved on the results page. For questions that call for comparison, verification and trust, people still feel the need to go to the source. A student’s question “what is academic coaching” ends in the summary; a parent’s decision about which course to enrol their child in does not.

The same experiment carries a warning that matters for businesses. The authors note that chat-based tools may push people to over-rely on the answer and may repeat misinformation⁸. So the summary is fast and comfortable, but not always right. When incomplete or outdated information about a product makes it into a summary, the user often does not check it by visiting the site; they settle for what they read and make their decision.

What This Means For A Business

Taken together, the evidence says this: on informational searches, when a summary sits at the top of the page, clicks to websites go down. How far they fall varies by measurement, sector and period; the direction is the same across every independent study³⁴⁵. What Google says about total click volume does not contradict that direction; it simply looks from a different scale⁷.

This does not mean SEO has lost its value. The sources that summaries draw on are, in the end, pages on the web; expecting a page that never appears in search results to make it into the answer is not realistic. What has changed is how success should be read. For a B2B manufacturer, visitor numbers on blog posts may no longer be as meaningful an indicator; the real question is whether the buying team comes across the company’s name while doing its research.

This is exactly where GEO’s subject begins: being named accurately inside the answer, even when no click follows. For an e-commerce site, that means the store’s name appearing in the answer given to someone asking about a product category; for a software company, it means being on the list in a reply that sets out the alternatives. As clicks decline, the measure of visibility sits somewhere a little wider than that familiar column in the report.