- A featured snippet extracts one passage from one page algorithmically. An AI Overview is LLM-generated text (Gemini) synthesized across multiple sources. Ahrefs measured an average of 7.7 domains per query for AI Overviews and 9.2 for AI Mode, not the single source a snippet pulls from.
- Semrush's own co-occurrence tracking found featured snippets and AI Overviews appearing together on the same results page fell from 34% in March 2025 to 18% by November 2025, real evidence that Google is increasingly substituting Overviews for snippets on a growing share of queries, not treating them as a complementary pair.
- The widely-repeated claim that holding a featured snippet roughly doubles your odds of an AI Overview citation could not be traced to any disclosed study despite a genuine search effort. It's flagged here explicitly as unsourced, following this site's no-fabrication standard.
- Real click-tracking data, not self-report, shows a large and apparently worsening hit to click-through when an AI Overview appears: Pew Research's panel study found traditional-link clicks fall from 15% to 8% when a summary is shown; Ahrefs' 300,000-keyword analysis found top-ranking pages losing 58% of their clicks, up from a 34.5% decline measured roughly eight months earlier.
- Trigger-rate estimates for how often AI Overviews appear at all range from 18% (Pew's real-panel data, March 2025) to newer 2026 estimates as high as the mid-40s, depending entirely on which keyword sample and methodology a tracker uses. The disagreement itself is the honest finding, not a number to resolve by picking a favorite source.
Both sit at the top of a Google results page, both get called "position zero" informally, and the two get treated as interchangeable in a lot of AI-search advice. They aren't the same system, they don't behave the same way, and, per actual click-tracking data, they don't cost the same amount of organic traffic. Here's the real comparison.
What's the mechanical difference?
A featured snippet is Google's classic algorithmic extraction: a passage, list, or table pulled directly from a single ranking page and displayed above the normal results, with that one page credited as the source. An AI Overview is a fundamentally different mechanism: text generated by Google's Gemini model, built using a technique Google calls "query fan-out," where the original query is broken into several sub-queries, multiple searches run concurrently, and the results synthesized into one generated answer rather than lifted verbatim from a single source.
That difference shows up directly in source count. Where a snippet credits exactly one page, Ahrefs' analysis found AI Overviews draw on an average of 7.7 domains per query, and Google's separate AI Mode surface draws on an average of 9.2, both meaningfully more diffuse than the single-source model a snippet uses.
Are they increasingly substitutes, not a pair?
Semrush's own tracking (100,000 US desktop and 100,000 US mobile keywords, measured in July 2025) found the rate at which a featured snippet and an AI Overview appear together on the same results page fell from 34% in March 2025 to 18% by November 2025. That's a real, measured decline in co-occurrence over less than a year, and it points toward AI Overviews increasingly replacing snippets on a growing share of queries rather than sitting alongside them as a complementary pair, the opposite of the "they reward the same thing" framing that's common in SEO advice.
Which claim doesn't survive a source check?
One specific claim shows up across a striking number of GEO and AI-search blogs, almost word for word: holding a featured snippet roughly doubles your odds of also being cited in an AI Overview for the same query. Despite a real, repeated search effort, this could not be traced to any single named study with a disclosed sample size or methodology. It has the exact profile of the stats this site has flagged before: a number that's been copied from post to post long enough to sound established, without anyone tracing it back to where it actually came from.
What the evidence does support, from multiple independent Ahrefs analyses over time, is a more general and less dramatic finding: pages already ranking in the top 10 organically are more likely to be cited in an AI Overview. That's a real correlation between ranking generally and AI Overview citation, not evidence that a featured snippet specifically confers any additional edge beyond what a strong organic ranking already provides.
Do they cite differently, even from Google's own other AI surface?
Google's AI Overviews and its separate AI Mode aren't even consistent with each other. A large Ahrefs study (540,000 matched query pairs) found the two surfaces cite different sources for the identical query roughly 87% of the time, reaching a similar overall conclusion but drawing on an almost entirely different source set to get there. That's the same instability already covered on this site as a reason a single-snapshot visibility check tells you very little. It applies just as directly to comparing a snippet's stable single source against either AI surface's shifting one.
What actually happens to clicks?
This is where the real data is strongest, because it comes from behavioral tracking rather than self-reported survey answers. Pew Research Center tracked actual browsing behavior, not a survey, from 900 US adults across 68,879 real Google searches in March 2025. An AI summary appeared on 18% of all tracked searches. When one appeared, users clicked through to a traditional result link only 8% of the time, versus 15% when no summary was shown; clicking a link embedded inside the AI summary itself happened in just 1% of visits; and users ended their search session entirely, with no further click at all, 26% of the time after seeing an AI summary, versus 16% without one.
A separate Ahrefs analysis (300,000 keywords, using real Google Search Console click data, comparing a December 2023 pre-AI-Overview baseline against December 2025) found average click-through for top-ranking pages down 58% when an AI Overview appears, worse than the 34.5% decline Ahrefs had measured roughly eight months earlier. The effect is getting bigger over time, not stabilizing. Position #2 lost roughly half its clicks; even position #10 fell about 20%. A third source, Seer Interactive, reports declines in a similar 49-65% range, though its methodology wasn't independently confirmed here. Three different measurement approaches, converging on the same direction: a real, and apparently still-growing, reduction in click-through wherever an AI Overview shows up.
How often does this actually happen?
Trigger-rate estimates vary more than almost any other figure in this space, and the disagreement is itself informative rather than a gap to paper over with one confident number. Pew's real-panel tracking found AI summaries on 18% of searches in March 2025. Conductor's Q1 2026 analysis of 21.9 million searches found 25.11%. Other 2026 trackers report figures as high as the mid-40s. These aren't measuring the same thing precisely. Different keyword sample sets, different geographies, and different query-type mixes all move the number substantially, which is exactly why no two trackers agree. The honest framing isn't "AI Overviews appear on X% of searches." It's that the range itself, and its lack of consensus, is what the current data actually supports.
Frequently asked questions
What's the actual difference between an AI Overview and a featured snippet?
A featured snippet is an algorithmic extraction of a passage, list, or table from a single ranking page, sometimes called "position zero." An AI Overview is generated text synthesized by Google's Gemini model, typically drawing on multiple sources at once rather than lifting one verbatim excerpt from a single page.
Does ranking for a featured snippet help you get cited in an AI Overview?
This is one of the most widely repeated claims in AI-search advice, and it could not be traced to any disclosed, methodologically-sound study despite a real search effort to find one. What is actually supported by multiple Ahrefs analyses is a more general, less exciting finding: pages already ranking in the top 10 organically are more likely to get cited in an AI Overview, a correlation with ranking generally, not evidence that holding a featured snippet specifically confers any additional advantage.
Do AI Overviews reduce click-through rate more than featured snippets did?
Real tracking data says yes, and by a wide and apparently growing margin. Pew Research's panel study (actual browsing behavior from 900 US adults, not self-report) found users clicked a traditional result link only 8% of the time when an AI summary appeared, versus 15% without one. Ahrefs' separate analysis of 300,000 keywords found top-ranking pages lost 58% of their click-through when an AI Overview appeared, up from a 34.5% decline measured about eight months earlier. The effect appears to be worsening over time, not stabilizing.
How often do AI Overviews actually appear in search results?
There's no single reliable answer, and that's itself the honest finding. Different trackers using different keyword samples report meaningfully different numbers, from Pew's real-panel 18% in March 2025 up to newer 2026 estimates in the 25-48% range depending on methodology. Treat any single confident percentage, including ones in this post, with the same skepticism the site applies to any undisclosed-methodology AI-search statistic.
Sources referenced
- Pew Research Center: Google users are less likely to click on links when an AI summary appears (Jul 2025, 900 US adults, 68,879 tracked searches)
- Ahrefs: AI Overview trigger analysis (146M SERPs)
- Ahrefs: AI Overviews / AI Mode source-divergence study (Dec 2025, 540,000 query pairs)
This piece builds on the instability findings already covered in the five stages for winning in the AI search era. For what actually gets a page cited once an AI surface does draw on it, see how to write content that AI models can actually quote. For the source-selection mechanics behind each surface specifically, see how AI Overviews choose sources and AI Mode vs. AI Overviews.
