Google’s AI Overviews have quietly rewritten the rules of search visibility, and a new dataset from Semrush spells out exactly what it now takes to get cited. According to Semrush’s analysis of more than 10 million keywords, roughly 97% of AI Overview citations come from pages already sitting in the top 20 organic results, and pages ranking in position one are pulled into an Overview more than half the time it appears. That is not a minor tweak to search behavior. It is a structural filter that decides which sites get seen by an AI summary and which ones simply disappear from view, even if they rank respectably on page one.
The timing matters. Semrush’s tracking shows AI Overview prevalence swung from 6.49% of queries in January 2025 to a peak of 24.61% in July 2025, before pulling back to roughly 15.69% by November 2025, according to Semrush’s December 2025 report. That volatility, paired with a sharp expansion into commercial search terms, is why publishers, SEO agencies, and product marketers are treating this Semrush update as the most consequential search-visibility story of the year. Tech Insider covered the click-through-rate collapse tied to AI Overviews in an earlier report; this analysis digs into the citation mechanics Semrush says now govern who gets quoted at all.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What Semrush’s Latest AI Overviews Data Actually Shows
Semrush’s study, built from a panel of more than 10 million tracked keywords, is the largest independent dataset available on how Google’s AI Overviews select and cite sources. The headline finding is straightforward: citation inside an AI Overview correlates almost entirely with existing organic ranking strength. Google is not running a separate, parallel ranking system for its AI summaries. It is drawing directly from the same top-20 pool it already ranks for the query, then compressing that pool down to a handful of citations.
That single data point reframes a debate that has consumed the SEO industry for two years. Many agencies pitched “AI Overview optimization” as a distinct discipline requiring new tactics, new schema, and new content formats. Semrush’s numbers suggest the opposite: the fastest way to appear in an AI Overview is still to rank in the traditional top 20, and ideally the top three, for the query in question. Google AI Overviews are not inventing a new visibility layer so much as re-packaging the existing one and, in the process, taking away the click.
Semrush also found that AI Overviews skew toward mid-difficulty, lower-volume queries rather than the biggest head terms. Nearly 60% of keywords triggering an AI Overview carry 100 or fewer monthly searches, and about 60% fall into the 21-60 Keyword Difficulty band. That is a meaningful signal for content teams: the queries most likely to produce an AI Overview are not the trophy keywords marketing teams chase, they are the long-tail, specific-intent searches buried deeper in a content calendar.
The Top-20 Rule: Why Ranking Position Still Decides Everything
The 97% figure is worth sitting with. It means that, for practical purposes, a page with zero presence in the top 20 organic results has almost no chance of being cited inside an AI Overview, regardless of how well-structured or comprehensive that page might be. Content quality and structure still matter, but they matter downstream of the ranking gate, not instead of it.
This creates a two-tier consequence for site owners. Sites already ranking well get an additional exposure channel, since an AI Overview citation can appear above the traditional blue links and carry a visible source attribution. Sites outside the top 20 get nothing extra to chase, because the AI layer simply is not built to look past that threshold. Semrush’s guidance for SEO teams follows directly from this: prioritize traditional ranking factors, including on-page optimization, internal linking, topical depth, and backlink acquisition, before spending budget on anything marketed specifically as “AI Overview SEO.”
Position One Now Captures More Than Half of All Citations
Beyond the top-20 filter, Semrush’s data shows an even steeper concentration at the very top. Pages holding the number-one organic position are cited in AI Overviews more than 50% of the time an Overview appears for that query. That is a far tighter concentration of visibility than traditional organic search, where position one typically earns a large but not majority share of clicks.
The practical effect is that the gap between ranking first and ranking fifth has widened. In classic organic search, a page in position five still captures meaningful traffic. Inside an AI Overview citation set, that same page’s odds of being quoted drop sharply the further it sits from the top slot. Search Engine Land’s review of the same Semrush dataset described 2025 as a year of AI Overview surge and pullback, noting that prevalence peaked at just under 25% of queries in July before Google visibly throttled the feature’s rollout later in the year.
AI Overviews Are Moving Into Commercial and Transactional Queries
The second major shift in Semrush’s report has nothing to do with ranking position and everything to do with query type. When Google first rolled out AI Overviews broadly, the feature concentrated on informational searches, the kind of “how does X work” queries that rarely carried commercial value. Semrush’s newer data shows that by late 2025 and into early 2026, roughly 42.9% of AI Overview queries carried commercial, transactional, or navigational intent, according to the company’s February 2026 analysis of ads and AI Overviews.
That is a fundamentally different risk profile for ecommerce brands, SaaS vendors, and any business whose revenue depends on organic search traffic for product and comparison queries. An AI Overview summarizing “best noise-canceling headphones under $200” or “cheapest project management software for small teams” is competing directly with product pages, comparison guides, and affiliate content that previously converted at meaningful rates. When Google’s own summary answers the question before a shopper clicks through, the commercial page underneath loses the chance to make its case.
Ads and AI Overviews Are Now Sharing the Same Results Page
Semrush’s most striking market-impact figure is not about organic citations at all. It is about advertising. The company found that search results pages showing both paid ads and an AI Overview together grew by more than 394% over the course of 2025. That is not a modest increase in an edge case; it describes a search results page that has been fundamentally restructured, with paid placements, an AI-generated summary, and organic results all now competing for the same limited screen space above the fold.
For advertisers, that convergence cuts two ways. Some analyses summarized by industry trackers have found that being cited within an AI Overview can coincide with meaningfully higher click-through on both organic and paid listings for the same query, since users who scroll past the AI summary tend to be higher-intent. Other data shows the opposite for brands that fail to appear anywhere in the Overview: their paid campaigns now compete for a shrinking pool of clicks that never leave the AI-generated summary at all. Either way, the SERP that PPC teams have optimized for over the past decade no longer resembles the one showing up in search results today.
Which Industries Face the Steepest Disruption
Not every vertical experiences AI Overviews the same way. Because the feature has expanded into commercial and comparison-heavy queries, industries built around product research and side-by-side evaluation are seeing the highest exposure rates. Retail, travel, software and B2B technology, and local services all show elevated AI Overview prevalence compared to purely informational or highly niche technical queries, based on comparative tracking cited alongside Semrush’s figures.
The common thread across these sectors is that they rely on comparison content, exactly the format Google’s AI systems are best equipped to summarize. A page that lists five competing products with prices and features is, structurally, already close to what an AI Overview wants to extract and repackage. That makes review sites, buying guides, and comparison articles some of the highest-value citation targets, but also some of the most exposed to having their core value proposition, “we compared these options so you don’t have to,” absorbed directly into the search results page itself. Sites publishing head-to-head guides, from AI assistant comparisons to product roundups, sit squarely in this exposed category.
The Content Formats Google’s AI Actually Prefers
Semrush’s broader research into AI-driven search, drawn from an analysis of 80 million AI search queries, found that listicle and how-to formats account for more than 40% of content cited across large language model-powered search tools generally. While that figure spans AI search platforms rather than AI Overviews specifically, it lines up with what Semrush’s Overview-specific data implies: structured, scannable content with clear headings, numbered steps, and direct comparisons is easier for an AI system to lift and summarize than long, narrative prose.
This is a case where good SEO practice and good AI-citation practice genuinely overlap. Content built around clear question-and-answer structure, explicit lists, and comparison tables tends to perform better in both traditional featured snippets and AI Overview citations. Below is a simplified example of the kind of structured markup that helps search systems parse a comparison page cleanly.
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [{
"@type": "Question",
"name": "What share of AI Overview citations come from top-20 pages?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Semrush's analysis found roughly 97% of citations come from pages already ranking in the top 20 organic results."
}
}]
}
</script>Semrush’s public guidance does not attach a specific numeric lift to schema markup itself. The recommendation is more general: structure content so headings, lists, and direct answers are easy for both human readers and automated summarization systems to extract without ambiguity.
How This Compares to Ahrefs’ Click-Through Rate Findings
Semrush’s citation-rule data answers who gets quoted. A separate line of research from Ahrefs answers what happens to everyone else’s traffic once an Overview appears. Ahrefs’ original study, covering 300,000 keywords, found that the presence of an AI Overview correlated with a 34.5% lower average click-through rate for the top-ranking organic page, with position-one CTR falling from 7.3% in March 2024 to 2.6% in March 2025 on desktop, according to Ahrefs’ April 2025 report.
By December 2025, that decline had nearly doubled. A follow-up Ahrefs study using Google Search Console data, again drawing on 300,000 keywords, found AI Overviews now correlate with a 58% lower average click-through rate for the top-ranking page compared with a pre-AI-Overview baseline, as detailed in Ahrefs’ February 2026 update. Read alongside Semrush’s citation data, the picture becomes clearer: the same top-20 pages that Google draws citations from are the ones absorbing the steepest traffic losses, because the citation itself often satisfies the user’s query before they ever reach the underlying page.
Semrush vs. Ahrefs vs. BrightEdge vs. Conductor: Whose Numbers Should You Trust
Four of the industry’s most-cited trackers have published overlapping but not identical figures for AI Overview prevalence, and the differences come down to methodology and query sampling rather than any one source being wrong. Semrush’s 10-million-keyword panel measured a peak of 24.61% prevalence in July 2025, dropping to 15.69% by November 2025. BrightEdge’s commercial-focused, nine-industry tracker recorded a much higher 48% prevalence, reflecting its concentration on commercial verticals where AI Overviews now appear more often. Conductor’s Q1 2026 benchmark, based on 21.9 million queries, landed at 25.11%.
| Tracking Source | Methodology | Reported AI Overview Prevalence | Measurement Period |
|---|---|---|---|
| Semrush | 10M+ keyword panel, mixed intent | 24.61% peak (Jul 2025), 15.69% (Nov 2025) | Jan 2025 – Nov 2025 |
| BrightEdge | 9-industry commercial tracker | ~48% | 2025-2026 rolling |
| Conductor | 21.9M query benchmark | 25.11% | Q1 2026 |
| Ahrefs | 300,000-keyword CTR panel | Not a prevalence tracker; CTR-focused | Mar 2025 – Dec 2025 |
The takeaway for anyone reading these reports side by side is not to pick a single “correct” number. It is to recognize that prevalence varies enormously by industry and query type, which is exactly why BrightEdge’s commercial-weighted sample lands so much higher than Semrush’s broader, mixed-intent panel. A business operating in a heavily commercial vertical should assume its real-world exposure sits closer to the BrightEdge figure than the more conservative Semrush average.
A Year of Surge and Pullback: The 2025 AI Overview Timeline
Semrush’s month-by-month tracking tells a story of rapid expansion followed by a deliberate correction. AI Overviews appeared on just 6.49% of tracked keywords in January 2025. By March 2025, that figure had more than doubled to 13.14% of U.S. desktop queries, a 102% jump in two months, according to a joint Semrush and Datos analysis of more than 10 million keywords. Growth continued through the spring and summer, peaking at 24.61% in July 2025, before Google appears to have pulled back the feature’s reach, with prevalence falling to roughly 15.69% by November 2025.
| Month | AI Overview Prevalence | Notable Shift |
|---|---|---|
| January 2025 | 6.49% | Baseline measurement |
| March 2025 | 13.14% | 102% increase in two months |
| July 2025 | 24.61% | Peak prevalence for the year |
| November 2025 | 15.69% | Pullback from summer peak |
| Late 2025 / Early 2026 | 42.9% of AIO queries | Share carrying commercial or transactional intent |
That pullback does not mean AI Overviews are fading. It more likely reflects Google tuning the feature toward queries where it performs best and away from ones where it either failed to add value or drew visible backlash from publishers and users. The parallel rise in commercial-intent coverage suggests Google shifted its remaining AI Overview real estate toward higher-value queries even as overall raw prevalence declined.
Historical Context: From Featured Snippets to AI Overviews
None of this is entirely new territory for the SEO industry, and it echoes debates already playing out around Google’s broader AI Mode rollout. Google’s featured snippets, introduced roughly a decade ago, triggered similar anxiety about “answer boxes” stealing clicks from the pages they were sourced from. Position Zero, as it came to be known, rewarded concise, well-structured answers and created its own optimization discipline. AI Overviews extend that same logic several steps further: instead of quoting a single snippet from a single page, Google now synthesizes an answer from several top-ranking sources at once, with citations attached.
The difference in scale is what makes this moment distinct. Featured snippets applied to a relatively narrow slice of queries. AI Overviews, even after their late-2025 pullback, still touch somewhere between 15% and 25% of tracked search queries depending on whose methodology is used, and closer to half of queries in commercially dense verticals per BrightEdge’s tracking. That is not a niche SERP feature anymore. It is a mainstream layer of how a meaningful share of all Google searches get answered.
Market Impact: Winners, Losers, and the Zero-Click Economy
The combined effect of Semrush’s citation data and Ahrefs’ CTR figures points toward an accelerating “zero-click” search economy, where a growing share of queries get fully answered without the user ever visiting a publisher’s site. Industry synthesis covering twelve separate studies on AI Overview impact found that every study measuring click-through rate found a decline, with the range spanning from 15% to as much as 89% depending on query type and measurement window. The same synthesis noted that pages actually cited within an AI Overview saw roughly 35% more organic clicks and 91% more paid clicks than uncited competitors for the same query, suggesting a winner-takes-more dynamic rather than a uniform loss across the board.
That divide is reshaping how publishers think about ad revenue and affiliate income. A site that consistently earns citations is capturing a disproportionate share of whatever traffic remains, while a site that ranks respectably but never gets quoted is losing ground on both organic clicks and, increasingly, paid visibility as ads and AI Overviews now co-occur on the same results page far more often than they did in 2024. Media companies and product-review publishers, many of which built entire business models around ranking in the top 10 for comparison queries, are the segment facing the most direct exposure to this shift, a dynamic also visible in how AI search tools compete for the same queries outside Google’s own results page.
What SEO Teams and Publishers Should Do Right Now
Semrush’s own recommendations, drawn directly from its citation data, boil down to a few concrete priorities. First, treat top-20 and top-three organic ranking as the primary lever, since 97% of citations never look past that pool. Second, build out topical depth around mid-difficulty, lower-volume queries, since those are disproportionately represented among AI Overview triggers. Third, structure content in clearly extractable formats, lists, defined steps, comparison tables, since those formats already dominate what large language model-powered search tools cite. Fourth, monitor paid campaign performance separately from organic, since the 394% growth in ad-plus-AI-Overview SERPs means paid budgets now need to account for AI-generated context sitting directly above or beside them.
None of these recommendations require abandoning fundamental SEO discipline. If anything, Semrush’s data is an argument against the cottage industry of AI Overview shortcuts that emerged over the past two years. The pages winning citations are, overwhelmingly, the pages that were already winning organic rankings through conventional means: authority, relevance, and structure.
Predictions: Where AI Overviews Go From Here
- Commercial and transactional query coverage will keep climbing past the current 42.9% share, as Google leans further into product and comparison summaries where advertising revenue and AI Overviews can coexist.
- The ads-plus-AI-Overview SERP layout, already up 394% in 2025, will likely become the default results page format for high-value commercial searches rather than an edge case.
- Expect continued volatility in raw prevalence figures, similar to the swing from 24.61% in July 2025 to 15.69% in November, as Google keeps tuning which query types justify an AI Overview.
- CTR pressure on top-ranking pages will likely deepen further past the current 58% decline reported by Ahrefs, particularly for informational queries where an AI summary can fully satisfy user intent.
- Publishers with strong topical authority in mid-difficulty, long-tail clusters stand to gain disproportionate citation share, widening the gap between sites that consistently get quoted and those that never do.
The Bottom Line for Businesses Depending on Organic Search
Semrush’s data closes the loop on a question that has hung over the SEO industry since AI Overviews launched broadly: is there a separate playbook for winning AI citations, or does traditional ranking still decide the outcome. The answer, based on the 97% top-20 citation rate and the more-than-50% position-one dominance, is that traditional ranking still decides almost everything. What has changed is the reward structure sitting on top of that ranking. A page that wins the traditional battle for position one now also wins a disproportionate share of AI citations, while everyone else competes for a shrinking pool of clicks in a search results page increasingly dominated by ads and AI-generated summaries.
For an industry that spent much of 2025 debating whether SEO was becoming obsolete, Semrush’s numbers argue the opposite. SEO fundamentals matter more, not less, because the AI layer amplifies whatever advantage a page already has in traditional rankings. The businesses most exposed are not the ones doing SEO wrong. They are the ones sitting just outside the top 20, close enough to see the opportunity and far enough away to never benefit from it.
Frequently Asked Questions
What percentage of AI Overview citations come from top-ranking pages?
Semrush’s analysis of more than 10 million keywords found that roughly 97% of AI Overview citations come from pages already ranking in the top 20 organic results for the query.
Does ranking number one guarantee an AI Overview citation?
Not a guarantee, but a strong advantage. Semrush found pages in position one are cited more than 50% of the time an AI Overview appears for that query, well above lower-ranking pages.
How much has Google AI Overviews reduced website traffic?
Ahrefs’ February 2026 update found AI Overviews now correlate with a 58% lower average click-through rate for the top-ranking page, up from a 34.5% decline measured in its April 2025 study, based on Google Search Console data across 300,000 keywords.
What share of search queries now show an AI Overview?
Estimates vary by tracker. Semrush measured a peak of 24.61% of queries in July 2025, pulling back to 15.69% by November 2025. BrightEdge’s commercial-focused tracker recorded a higher 48%, while Conductor’s Q1 2026 benchmark found 25.11% across 21.9 million queries.
Are AI Overviews mostly appearing on informational or commercial searches?
AI Overviews started out concentrated on informational queries but have expanded significantly. Semrush found that roughly 42.9% of AI Overview queries now carry commercial, transactional, or navigational intent, up from a much smaller share in early 2025.
What content formats does Google’s AI system prefer to cite?
Semrush’s broader research into AI-driven search found that listicle and how-to formats account for more than 40% of content cited across large language model-powered search tools, suggesting structured, scannable formats are favored over long narrative prose.
How does this compare to older Google features like featured snippets?
Featured snippets pulled a single answer from one page for a narrow slice of queries. AI Overviews synthesize answers from several top-ranking sources at once and now touch between roughly 15% and 48% of queries depending on the tracker and industry, making the effect far broader in scale.
Should businesses build a separate strategy just for AI Overviews?
Semrush’s data suggests no. Because 97% of citations come from pages already ranking in the top 20, the most effective strategy remains traditional SEO fundamentals, on-page optimization, topical depth, and authority building, rather than tactics marketed specifically as AI Overview optimization.
