We build therapist websites and sell visibility plans, so we have an interest in what AI Overviews do to clicks. To keep this guide fair, every figure comes from the study or from Google with its period and method, we say where each study falls short, Google’s statements are labeled as Google’s, and we don’t forecast. This guide goes deeper on one part of our SEO for therapists guide.

What are AI Overviews and AI Mode?

They are Google features that show an AI-written answer with links to supporting pages. Google says AI Overviews help “people get to the gist of a complicated topic or question more quickly,” and AI Mode is “particularly helpful for queries where further exploration, reasoning, or complex comparisons are needed,” returning “a comprehensive AI-powered response with links to supporting websites.”4

AI Overviews don’t appear on every search. Google says they are “only shown when our systems determine that it is additive to classic Search, and as such, often don’t trigger.”4

What do the studies show at a glance?

Three studies measured clicks. Their periods, methods and limits differ, so read across each row before comparing numbers.

StudyPeriod and methodWhat it foundMain limits
Pew Research Center, July 20251March 2025. Browsing data from 900 US adults, 68,879 Google searches.Result links clicked in 8% of visits with an AI summary and 15% of visits without.Observational, and Google searches only.
Agarwal and Sen, working paper, June 17, 2026 version6January 7 to February 10, 2026. Randomized. 1,065 US-based participants, desktop Chrome.On searches with an AI Overview, outbound organic clicks 39.8% lower.Working paper, and we found no journal version. One recruitment panel.
Wang and colleagues, preprint, August 20265March 2026. Randomized and preregistered. 1,100 US adults on Chrome, 7 days.For people whose AI Overviews were actually hidden, click-through rate 8.8 points higher (95% confidence interval 2.3 to 15.3). Comparing everyone assigned to hiding: +2.7 points, not statistically significant (p = 0.124).Preprint, and we found no journal version. Hiding failed on about half of AI Overviews.

What did Pew Research Center find?

In March 2025, Google users clicked a regular result link in 8% of visits to results pages with an AI summary and in 15% of visits to pages without one.1 Pew analyzed browsing data from 900 US adults: 68,879 unique Google searches, of which 12,593 produced an AI summary. People clicked a link inside the summary in 1% of visits to pages with one, and ended their browsing session on 26% of pages with a summary against 16% of pages without.

Two cautions from Pew itself. It re-ran each search between April 7 and 17, 2025 to record what the page showed, and says “The AI summary for any given search may change over time.”1 And with or without a summary, around two-thirds of all searches ended with the person browsing elsewhere on Google or leaving the site entirely, without clicking a result link.1 In our reading, because the study compares visits, the searches that got a summary may differ from those that didn’t in other ways.

What did the randomized experiments find?

Two experiments hid AI Overviews for some participants at random. Both found people clicked more when the summaries were hidden, though the second found its simplest comparison was not statistically significant.

Saharsh Agarwal (Indian School of Business) and Ananya Sen (Carnegie Mellon University)2 recruited 1,065 US-based participants through Prolific. A desktop Chrome extension hid AI Overviews for one group while a control group saw them as usual. Each participant was observed for two weeks, and data collection ran from January 7 to February 10, 2026. Their abstract says: “Conditional on appearing, AIOs reduce outbound organic clicks by 39.8% and increase zero-click searches by 34.5%, without affecting sponsored clicks or overall search frequency.” (“AIOs” are AI Overviews.) It adds that “Engagement on downstream websites, conditional on clicks, is not different across groups.” AI Overviews appeared on about 41% of queries. Across all searches, not only those with an AI Overview, the authors put the average effect at roughly 18.5% fewer outbound organic clicks and roughly 20.4% more zero-click searches.6 Every Agarwal and Sen figure in this guide is from that June 17, 2026 version, which we read in full. SSRN lists a later revision dated July 8, 2026, whose abstract gives the same 39.8% and 34.5% figures.2 We could not open its full text, so its other figures may differ.

Wang and colleagues ran a preregistered experiment with 1,100 US adults using Chrome, recruited March 17 to 19, 2026 and followed for seven days. AI Overviews appeared on 36% of searches. A Google interface change broke the extension’s hiding, and the authors report that compliance in that condition fell to roughly 50% overall. So they report two estimates of click-through rate (external clicks divided by Google searches). The first, for people whose AI Overviews were actually hidden, is an 8.8 percentage point rise, with a 95% confidence interval of 2.3 to 15.3. The authors note that it relies on an assumption that may not hold if the interface change altered participants’ experience more broadly. The second, the simpler comparison of everyone assigned to hiding, is +2.7 points, which is not statistically significant (p = 0.124). A third group was assigned to AI Mode only, and its click-through rate was 18.8 points lower than the group that searched as usual.5 Agarwal and Sen also report an exploratory AI Mode condition that suggests even greater reductions.6

Why do the studies’ numbers differ?

They agree on direction but measure different things, in different groups, at different times.

What does Google say about clicks?

Google says total click volume to websites has been stable, and it publishes no figures. Liz Reid, head of Google Search, wrote on August 6, 2025: “Overall, total organic click volume from Google Search to websites has been relatively stable year-over-year.” The same post says “people may be satisfied with the initial response and not click further” for some quick questions, and that shifting trends are “resulting in decreased traffic to some sites and increased traffic to others.”3 That post was written after Pew’s study and before both experiments.

Google’s AI features page says: “We’ve seen that when people click from search results pages with AI Overviews, these clicks are higher quality (meaning, users are more likely to spend more time on the site).”4 The Agarwal and Sen experiment reports no difference in engagement on websites after a click. The two measure different things, and Google’s page gives no figures.

Which therapy searches are most exposed?

No study measured therapy searches, so this table is our inference from how often each kind of search showed an AI summary. The example searches are ours.

Kind of searchExampleWhat the studies showOur inference
Long, question-style searches“how do I know if we need couples counseling”Pew: 60% of searches starting with who, what, when or why showed a summary, and 53% of searches of 10 or more words.1 Agarwal and Sen: informational searches showed an AI Overview 53% of the time, and hiding it had “a large and positive effect” on clicks.6Most exposed.
Short searches“therapist Pasadena”Pew: 8% of one- or two-word searches showed a summary.1Less exposed. No study separated local searches.
Your practice name“[practice name] Pasadena”Agarwal and Sen: navigational searches showed an AI Overview 15% of the time, and hiding it had “statistically insignificant effects” on clicks, from a smaller sample.6Less exposed.
Specific decision searches“couples therapist in [city] who takes [insurance]”Not measured. The closest category, transactional searches, showed an AI Overview 6% of the time, and hiding it had statistically insignificant effects.6Unknown.

The authors sorted searches into these categories with an AI model, so the labels are approximate. Wang and colleagues say future work should measure “specific high-stakes contexts and search tasks (e.g., health, legal, or financial information-seeking).”5

How can I see AI features in my own Search Console data?

Open the Generative AI performance report, then compare it with your regular Performance report over the same months.

What should I measure?

MeasureWhereWhy
Inquiries per month (calls, forms, bookings)Your own scheduler, email and phone recordsIt is the outcome you care about, and a click is one step toward it. This is our judgment.
Branded clicks and impressionsPerformance report, branded filterGoogle says “Branded traffic often comes from users already familiar with your site.”8
Non-branded clicks and impressionsPerformance report, non-branded filterGoogle says “non-branded traffic can indicate new audience growth.”8
AI feature impressions by pageGenerative AI performance report, Pages tabShows which pages Google links in AI Overviews and AI Mode. Impressions only.7
A dated note of your site changesYour own logLets you line up a change in the numbers with what you changed.

What should a practice website publish in response?

Google says unique, useful content is the part most likely to matter. Its AI guide says: “Creating content that people find unique, compelling, and useful will likely influence your website’s presence in generative AI search in the long run more than any of the other suggestions in this guide.”10

For a practice, our reading is pages only you can write: who you see, where, how sessions work, what they cost and how to start. Those answer the specific searches in the table above, where AI summaries appeared less often. Our guides on listing fees, local SEO and getting mentioned by AI assistants cover them. We don’t forecast what AI Overviews will do to clicks, and we found no measurement that would support a forecast.

Questions therapists also ask

Should I stop writing blog posts about symptoms and concerns?

This is our judgment: nothing in these studies says to stop. They suggest long, question-style searches are the ones most likely to show an AI summary, so decide what each article is for and look at the inquiries it leads to in your own records. Our guide on what Google says about AI-written content covers the quality bar.

Can I opt my site out of AI Overviews?

Yes. In Search Console, the Search generative AI control lets you include your site (the default), exclude it, or inherit the setting of a parent property. If you exclude it, Google says “You won’t receive any traffic or impressions from these features,” and that the control “isn’t used as a ranking or inclusion signal affecting other parts of Search.” It generally takes a few days to apply.11 We found no study of what excluding does to a site’s visits, so we don’t recommend it either way.

Are clicks from AI Overviews better?

Google says they are higher quality, as quoted above. The Agarwal and Sen experiment found no difference in engagement on downstream websites after a click, and Google’s page gives no figures, so we can’t tell you which is right for therapy sites.

Do AI Overviews affect searches for a therapist in my city?

We found no study that measured them. Pew found that 8% of one- or two-word searches produced a summary, which suggests short local searches are less exposed than long questions, but Pew did not separate local searches. Check your own Search Console data rather than assuming.

Will this get worse?

We can’t say. These studies are snapshots from March 2025 to March 2026, and Google says AI Overviews “often don’t trigger.” A monthly comparison of your own numbers will tell you more about your site than a forecast.

Where to start

Write down your inquiries by month from your own records, then add two Search Console views: branded against non-branded clicks, and the Generative AI performance report’s impressions by page. Compare the same months. If you would like us to do it, our Visibility plan ($349 a month, or $279 a month billed annually) includes a monthly report with “available search and inquiry observations,” and AI mentions and citations each month, with sources and limits. We don’t promise rankings, traffic or AI recommendations. Find the plan that fits your practice.

To see what our monthly search and AI visibility work includes, see search and AI visibility for therapists.

This guide reports studies, two of them working papers or preprints that we found no journal version of, and Google’s own statements. Google’s products change often, so check the linked pages before you act.