AI Overviews Study: What 7 Major Datasets Actually Tell Us (And Where They Disagree)
TL;DR
- Seven major AI Overviews studies from Semrush, Seer Interactive, Pew Research, BrightEdge, Conductor, seoClarity, and Profound collectively analyzed over 130 million keywords and 680 million citations, and they don’t agree on how bad the damage really is.
- Organic CTR for queries with AI Overviews dropped 61% year-over-year according to Seer Interactive’s September 2025 study, but Semrush’s 200K-keyword analysis found that zero-click rates actually decreased slightly when AIOs appeared on previously AIO-free queries.
- The real strategic move isn’t panicking or ignoring AI Overviews. It’s using the AIO Triage Matrix in this article to figure out which data applies to YOUR business, then acting on that, not on headlines.
The studies can’t agree. That’s actually the story.
I spent the last three months reading every major AI Overviews study I could find. And here’s the thing nobody’s saying out loud: they contradict each other. A lot.
One study says AI Overviews are barely expanding. Another says they’ve surged nearly 475% year-over-year. One dataset shows zero-click rates going down after AIOs appear. A completely different study says users end their browsing session 63% more often when they see an AI summary. Both cite real numbers. Both used large samples.
If you’ve been reading these studies one at a time and walking away with a clean narrative (“AI Overviews are killing traffic” or “AI Overviews aren’t that bad”), you’ve been reading them wrong. The truth lives in the gaps between them, and that’s exactly where we’re going.
By the end of this piece, you’ll have a clear picture of what’s actually happening with AI Overviews based on cross-referencing seven major datasets, plus an original framework for deciding which findings matter for your specific business. No cherry-picking. No doom-mongering. Just the data, the conflicts, and what to do about them.
How big is the AI Overviews footprint, really?
Depends on who you ask. And I don’t mean that as a cop-out.
Semrush’s 10-million-keyword study (updated December 2025) found AI Overviews appearing on 15.69% of queries in November 2025, down from a peak of 24.61% in July 2025. That’s a meaningful pullback. Meanwhile, Conductor’s 118-million-keyword analysis pegged AIO prevalence at 16% in September 2025, showing an 88% increase since April of that year. And seoClarity reported a staggering 475% year-over-year surge in AIO presence for mobile U.S. keywords between September 2024 and September 2025, with AIOs appearing on 30% of desktop keywords.
How do you get 15.69%, 16%, and 30% from studies conducted within weeks of each other? Methodology. Semrush analyzed a broader keyword set including branded and navigational terms. Conductor tracked a different 118 million keywords. seoClarity focused specifically on desktop keywords in their Research Grid, which skews toward more competitive terms.
| Study | Keywords Analyzed | AIO Prevalence | Time Period | Key Caveat |
|---|---|---|---|---|
| Semrush | 10M+ | 15.69% (Nov 2025) | Jan–Nov 2025 | Includes all query types |
| Conductor | 118M | 16% (Sep 2025) | Monthly, ongoing | Desktop leads mobile |
| seoClarity | Not disclosed | 30% desktop (Sep 2025) | Sep 2024–Sep 2025 | Desktop-only figure |
| BrightEdge | Not disclosed | 11%+ (May 2025) | May 2024–May 2025 | Enterprise client base |
| McKinsey | N/A (survey-based) | ~50% of searches show AI summaries | Oct 2025 report | Broader definition includes all AI features |
McKinsey’s October 2025 report claimed about 50% of Google searches already feature AI summaries, with that number expected to reach 75% by 2028. That’s wildly higher than the SEO tool data. The likely explanation? McKinsey used a broader definition of “AI summaries” that includes features beyond the strict AI Overview box, and they surveyed consumer behavior rather than scraping SERPs.
Pro Tip: When you see an AI Overviews stat, always check three things before acting on it: the keyword sample size, whether the data covers mobile or desktop (or both), and what the study counts as an “AI Overview.” These three variables explain most of the discrepancies between studies.
The honest answer? AI Overviews probably appear on somewhere between 15% and 30% of queries depending on device, industry, and query type. That range isn’t a failure of research. It’s a reflection of how volatile Google’s AIO rollout has been.
Are AI Overviews actually stealing clicks? The data is weirder than you think.
Here’s where things get genuinely confusing, and where I think most marketers are getting the story wrong.
Seer Interactive tracked 3,119 search terms across 42 client organizations from June 2024 through September 2025 and found organic CTR for queries with AI Overviews plummeted 61%, dropping from 1.76% to 0.61%. Paid CTR crashed even harder: 68% decline, from 19.70% down to 6.34%. Those are brutal numbers.
But Semrush, using a different approach, tracked over 200,000 keywords and compared the zero-click rate of queries before and after they started showing AI Overviews. Their finding? The zero-click rate actually decreased from 33.75% to 31.53% when AIOs appeared. That directly contradicts the “AIOs kill clicks” narrative.
And then there’s Pew Research. Their study of 900 U.S. adults’ actual browsing behavior in March 2025 found that users who encountered an AI summary clicked on a traditional search result only 8% of the time, versus 15% for users who didn’t see one. Users ended their browsing session entirely 26% of the time when an AI summary appeared, compared to 16% without one.
So which is it? Clicks going down, zero-click rates going down, or both?
All three. And they’re not actually contradictory once you understand the mechanics. Think of it like measuring rainfall in a city. One weather station tracks total rainfall per square mile. Another measures how many individual storms produce rain. A third asks residents whether they got wet. They’ll give you different numbers, but they’re all measuring real phenomena.
Seer’s data captures the rate at which people click when they see an AIO. That rate is tanking. Semrush’s data captures whether the type of query that gets an AIO was already low-click to begin with, and those queries were. The Pew data captures real user behavior at the session level, including what people do after they see the summary.
The reconciled picture looks something like this: Google is placing AIOs predominantly on queries that already had low click-through rates (informational, long-tail). When AIOs appear, users click even less often on individual results. But because those queries were already low-click, the incremental zero-click increase per query is modest. The aggregate traffic loss, however, is massive because AIOs are appearing on millions of queries simultaneously.
The scariest finding nobody’s talking about
Seer’s study contained a detail that should terrify anyone running informational content strategies. Even queries without AI Overviews saw organic CTR decline 41% year-over-year, dropping from 2.72% in June 2024 to 1.62% by September 2025.
That means CTR is declining across the board, not just on AIO queries. Users are finding answers somewhere, but increasingly not by clicking organic results. Whether that’s because of AI chat tools, social search, or simply because SERPs have gotten more answer-rich, the effect is the same: the old playbook of “rank and bank” is broken for informational content, AIO or no AIO.
Who’s actually getting cited? The citation game is rigged (but predictable).
If clicks are drying up, then getting cited inside the AI Overview becomes the new form of visibility. But who’s Google choosing to cite?
Profound’s analysis of 680 million citations across ChatGPT, Google AI Overviews, and Perplexity between August 2024 and June 2025 revealed that each platform plays favorites differently.
| Platform | #1 Cited Source | % of Citations | #2 Source | Key Pattern |
|---|---|---|---|---|
| ChatGPT | Wikipedia | 7.8% | Reddit (1.8%) | Encyclopedic authority bias |
| Google AI Overviews | 2.2% | YouTube (1.9%) | Social + video preference | |
| Perplexity | 6.6% | YouTube (2.0%) | Community-driven answers |
Reddit dominates Google AI Overviews and Perplexity. Wikipedia dominates ChatGPT. This isn’t random. Google’s AI Overviews are built to synthesize multiple perspectives, and Reddit threads naturally provide exactly that: multiple viewpoints, upvoted by real users, on specific questions.
Pew Research confirmed this pattern from the user side, finding that Wikipedia, YouTube, and Reddit collectively accounted for 15% of sources listed in AI summaries. Government websites (.gov) appeared in 6% of AI Overview citations versus just 2% in standard search results.
Here’s what this means practically: if you’re a B2B SaaS company, your blog post is competing for AIO citations against Reddit threads, YouTube explainers, and LinkedIn articles. The bar for citation isn’t just “good content.” It’s content that Google’s AI considers more synthesis-worthy than a Reddit thread with 47 upvoted replies.
“The assumption was that AI would kill traditional search. But our data tells a different story: search is not disappearing, it’s expanding.”
— Jim Yu, Founder and CEO of BrightEdge (BrightEdge One-Year AIO Report, May 2025)
That quote sounds optimistic, and it’s backed by BrightEdge’s finding that total search impressions increased over 49% since AI Overviews launched. But more impressions with fewer clicks per impression is cold comfort if your revenue model depends on website visits.
The AIO Triage Matrix: a framework for knowing what to do
Here’s what I haven’t seen in any other AI Overviews analysis: a practical decision tool. Every article gives you data. None of them help you figure out which data points matter for your situation. So I built one.
The AIO Triage Matrix asks three questions about your business. Your answers determine which studies you should weight most heavily and what to do next.
Question 1: What percentage of your organic traffic comes from informational queries?
If it’s above 50%, the Seer Interactive and Pew data should keep you up at night. That 61% CTR decline and the “26% of sessions end after seeing an AIO” stat hit your business model directly. Your move: start measuring AIO citation share as a primary KPI, not a nice-to-have.
If it’s below 30%, you’re in relatively better shape. Semrush’s data shows commercial and transactional queries saw AIO prevalence grow (commercial intent queries with AIOs increased from 8.15% to 18.57% over 2025), but the CTR impact on buying-intent queries is less severe than on informational ones.
Question 2: Are your top keywords in a high-AIO-saturation industry?
Semrush’s industry breakdown shows massive variation. Science (25.96% AIO saturation), Computers & Electronics (17.92%), and People & Society (17.29%) are getting hammered. Real Estate and Shopping? Under 3%.
If you’re in healthcare, education, B2B tech, or insurance (BrightEdge’s highest-AIO industries), you need to treat AIO optimization as equal to traditional SEO. If you’re in e-commerce or local services, traditional SEO still drives most of your traffic.
Question 3: Do you currently appear in AI Overviews for your target queries?
This is the make-or-break variable. Seer’s data shows that when your brand IS cited in an AI Overview, you get 35% higher organic CTR and 91% higher paid CTR compared to AIO queries where you’re not cited. Being inside the AIO is dramatically better than being pushed below it.
Watch Out: Seer Interactive explicitly cautions that they “cannot definitively prove that citation causes higher CTRs.” Brands with stronger authority may simply be more likely to get both citations and clicks. The correlation is strong, but don’t assume that gaming your way into an AIO citation will automatically boost your click rate.
Here’s the matrix in action:
- High informational traffic + high-AIO industry + not cited: Red alert. Shift 20-30% of content resources toward AIO optimization. Audit your top 50 informational pages for structured data, concise answer formatting, and topical authority signals.
- High informational traffic + high-AIO industry + already cited: You’re in the game. Focus on expanding citation coverage to more queries and monitoring for citation loss. Track citation share monthly.
- Low informational traffic + low-AIO industry + not cited: Watch and wait. Invest in traditional SEO for now, but build the infrastructure (structured data, clear answer formatting) so you’re ready when AIOs expand into your vertical.
- Mixed traffic + any industry + any citation status: Run the Seer-style analysis on your own data. Segment your queries by AIO presence in Google Search Console, then compare CTR trends. The aggregate studies are directionally useful, but your specific query portfolio is what matters.
What publishers are seeing on the ground
The aggregated data studies tell one story. Publisher-level reporting tells another, messier one.
Digital Content Next surveyed 19 of its member companies (which include the New York Times, Conde Nast, and Vox) between May and June 2025. The median year-over-year decline in Google Search referral traffic was 10% overall. News brands saw a 7% decline. Non-news brands saw 14%.
Those numbers sound almost manageable compared to the CTR studies. But individual cases tell a different story. The Professional Publishers Association submitted evidence to the UK’s Competition and Markets Authority showing one lifestyle publisher’s CTR on a popular query dropped from 5.1% to 0.6% over a single year, despite stable rankings and steady impressions.
That’s an 88% CTR collapse on a single keyword. And it happened while the page still ranked on page one.
An AIO (AI Overview) is a generative AI-produced summary that appears above traditional organic results on Google Search, pulling information from multiple web sources into a single synthesized answer.
The disconnect between “median 10% traffic decline” and “88% CTR collapse on individual queries” isn’t a contradiction. It’s a distribution problem. Some queries lose almost everything. Others are barely touched. The median hides the extremes, and if one of those extreme queries happens to drive 40% of your revenue, the median is irrelevant to you.
Where this is headed (and what I’d bet on)
I’ll be honest: I didn’t expect Semrush’s zero-click data to push back against the doom narrative as hard as it did. And I didn’t expect Seer’s “even non-AIO queries are declining” finding to be as stark.
Here’s what I’d bet on for the rest of 2026, based on what the cross-study data suggests:
AIO prevalence will stabilize between 20-35% of queries. Google pulled back from its July 2025 peak, which tells me they’re calibrating, not just expanding blindly. The shift toward commercial and transactional queries (Semrush saw transactional AIO queries jump from 1.98% to 13.94%) will continue, which means e-commerce and SaaS brands that feel safe today won’t feel safe for long.
CTR for informational queries will keep declining regardless of AIOs. Seer’s data on non-AIO queries declining 41% is the most underreported finding in any of these studies. The problem isn’t just AI Overviews. It’s a fundamental shift in how people seek information. ChatGPT, Perplexity, social search, and Google’s own answer-rich SERPs are all pulling clicks away from traditional results.
Citation will become a measurable, trackable KPI. Right now, most brands can’t systematically track whether they appear in AI Overviews. That will change fast. BrightEdge reports over 750 organizations already use its AI Catalyst tool for exactly this purpose.
The brands that win will be the ones that stop treating “AI Overviews” as a single monolithic threat and start treating it as a segmented, query-by-query challenge. Which is exactly what the AIO Triage Matrix is designed to help with.
If you want a team that can run this kind of analysis for your specific keyword portfolio and build a strategy around it, LoudScale does exactly that.
Frequently Asked Questions About AI Overviews Studies
How much do AI Overviews reduce organic click-through rates?
Seer Interactive’s September 2025 study of 3,119 search terms found organic CTR dropped 61% for queries with AI Overviews, from 1.76% to 0.61% year-over-year. However, Semrush’s analysis of 200,000+ keywords found that zero-click rates actually decreased slightly (from 33.75% to 31.53%) when AIOs appeared on previously AIO-free queries, suggesting the queries Google selects for AIOs were already low-click to begin with.
What percentage of Google searches trigger an AI Overview?
As of late 2025, the major studies place AIO prevalence between 11% and 30% depending on methodology. Semrush reported 15.69% of 10 million keywords in November 2025. Conductor found 16% of 118 million keywords in September 2025. seoClarity measured 30% of U.S. desktop keywords that same month. McKinsey’s broader consumer survey estimated about 50% of Google searches feature some form of AI summary.
Which websites get cited most often in Google AI Overviews?
Profound’s analysis of 680 million citations found Reddit (2.2% of total citations), YouTube (1.9%), Quora (1.5%), and LinkedIn (1.3%) are the top sources cited in Google AI Overviews. Pew Research separately confirmed that Wikipedia, YouTube, and Reddit collectively accounted for 15% of sources in AI summaries, with government (.gov) websites overrepresented at 6% of AIO citations versus 2% in standard results.
Does getting cited in an AI Overview actually help your traffic?
Seer Interactive’s data shows brands cited in AI Overviews get 35% higher organic CTR and 91% higher paid CTR compared to AIO queries where they aren’t cited. But Seer explicitly notes they cannot prove causation: brands with stronger authority signals may naturally earn both more citations and more clicks. The correlation is consistent across 15 months of data, but correlation isn’t causation.
Which industries are most affected by AI Overviews?
Semrush’s November 2025 data shows Science (25.96% of keywords triggering AIOs), Computers & Electronics (17.92%), and People & Society (17.29%) have the highest AIO saturation. BrightEdge identified Healthcare, Education, B2B Tech, and Insurance as the industries with the strongest AIO presence. Real Estate and Shopping consistently show the lowest AIO saturation at under 3% of keywords.