Google AI Overviews Reach 88% of Commercial Keywords in Semrush Panel

AI Overviews have moved into buying-intent searches

Google's AI Overviews were first seen mainly as a tool for informational searches like "how do I" and "what is" queries. A new dataset suggests that view is out of date. Semrush tracked 400 fixed U.S. desktop keywords in banking, ecommerce and software for a year. Over that time, the share that triggered an AI Overview rose from 45.75% in September 2025 to 88% in August 2026.12 The keywords were picked to match the last stage before a purchase. Each one carried a modifier such as "alternatives," "brand vs. brand," "best X for Y," "price/cost" or "reviews."12

Review searches showed the biggest change. A year earlier, review-modified queries triggered an AI Overview only about 30% of the time. By August the figure was 89.04%, a gain of almost 59 percentage points.13 "Best X for Y" queries ended with the highest coverage of any group, at 96.12%. Brand-versus-brand queries started highest at 63.41% and grew the least, ending at 86.59%.13 Every modifier group finished above 80%, so the gap between the groups narrowed a lot over the year.11

Weighting by search volume makes the figure higher still. When each keyword counts according to Semrush's estimate of monthly demand, AI Overviews showed up on 97.08% of commercial searches in the panel. The 12% of keywords that never triggered one made up less than 3% of estimated volume.12 Put simply, the searches buyers run most often were also the most likely to open with an AI-written answer.

Ecommerce caught up with banking and software

The gap between industries closed even faster than the gap between query types. In September 2025, coverage ran from 32.62% in ecommerce to 68% in software. By August, all three industries were above 85%, with software at 91%, banking at 89% and ecommerce at 85.11%.13 Three industry-and-modifier combinations reached full coverage: banking "best X for Y," software "alternatives" and software "best X for Y."12 Three stayed below 80%: banking alternatives at 65%, ecommerce price/cost at 71.43%, and software reviews at 75%.12

Ecommerce price queries changed the most dramatically. Coverage jumped from 4.76% to 71.43% in that segment. PPC Land worked out that this probably means a move from one keyword to fifteen out of about 21.13 That arithmetic points to the study's main weakness, discussed below.

The figures conflict with earlier studies, and both can hold

Earlier large-scale studies described a different picture. WebFX analyzed about 2.37 million keywords and found AI Overviews on roughly one in four U.S. searches overall, and on 22.2% of commercial-intent queries.15 Its finding was that brand and retailer modifiers acted as "brakes" on whether an Overview appeared.15 Another analysis of 300,000 keywords found that commercial queries made up only 5.8% of the results pages that showed an AI Overview.20

The sources don't really contradict each other. They measure different things. Semrush's panel is a small, hand-picked set of decision-stage queries, not a sample of all Google searches. Both Search Engine Land and an independent AI Overviews tracker say the 88% figure describes this sample and should not be read as how common the feature is overall.1119 The direction matches, though. Semrush's earlier study covered more than 600,000 keywords and found that commercial-intent results with an AI Overview grew 71% on average between November 2025 and April 2026.13 Our reading: AI Overviews have expanded quickly into commercial search over the past year. The exact 88% depends on how the panel was built, but the trend appears in more than one dataset.

The coverage also identified real gaps in the method. The study counts only whether an AI Overview appeared. It has no data on clicks, traffic or conversions, and it covers desktop only.13 PPC Land estimated that each industry-and-modifier cell held roughly 20 to 30 keywords, so a single keyword can move a cell by three to five points.13 PPC Land also took issue with the study's claim that "every searcher" saw an AI Overview in the cells at 100%, since the metric counts keywords, not people.13 It further pointed out that Semrush sells the visibility-tracking tools its own recommendations call for.14 None of this undercuts the overall trend, but it is a reason to treat the precise numbers with caution.

The traffic question

Because the Semrush study has no click data, it doesn't show what this shift costs websites. Other research does. Pew Research Center tracked about 68,000 real searches and found users clicked a result 8% of the time when an AI summary appeared, compared with 15% without one.1 Sessions also ended outright more often after an AI Overview: 26% versus 16%.3 Ahrefs' February 2026 update put the click-through loss for the top-ranking page at 58% when an Overview is present, up from 34.5% in its April 2025 analysis.2

The headline numbers are often quoted carelessly. One German analysis notes that the Ahrefs figure applies to the top-ranking page on keywords that trigger an Overview. It is not an average traffic loss across all sites.3 A critique of the Ahrefs data also found that click-through fell sharply on keywords with no AI Overview at all, which suggests part of the decline has other causes.6 Seer Interactive found that click-through on AI Overview queries hit a low of 1.3% in December 2025 and recovered to 2.4% by February 2026. That is still well below the roughly 3.3% seen when no Overview appears.7

The commercial angle complicates matters. First Page Sage reported that ecommerce and retail were among the least affected industries, losing 11.6% of organic sessions, because transactional searches still need a click to finish. Media and publishing lost 47.3%.4 The same report found that visitors who clicked through from pages with an AI Overview stayed longer and converted to leads at 2.6%, compared with 1.7%.4 Brand owners may lose fewer clicks than the coverage figures imply, and the clicks they keep may be worth more.

The outlook is worse for the publishers whose business depends on commercial queries: review sites, comparison sites and affiliate publishers. A "best X for Y" or "X review" search is exactly the kind of query those businesses were built on. That category now shows an AI Overview on about nine in ten tracked keywords.13 Chartbeat data shows search referrals fell 60% for small publishers over two years, 47% for mid-sized ones and 22% for large ones.5 Penske Media has sued Google on antitrust grounds, and the European Publishers Council has filed a complaint with the European Commission.2 Google, for its part, added features in May 2026 meant to send traffic back to sites, including inline links, hover previews and a "Further Exploration" section.3

Who gets cited matters as much as ranking

The most useful lesson from the study may be its example of how Google builds these answers. On a Klaviyo-versus-Mailchimp query, the AI Overview cited Klaviyo's own guide to argue for Klaviyo. To argue for Mailchimp, it cited Zapier, a third-party site, along with a YouTube video and another outside site.12 PPC Land described the result plainly: a brand's own page is now one source among several in the first answer a buyer sees.13 Semrush's June 2026 visibility report counted an average of 9.2 citations per AI Overview, and brands compete with reviewers, creators and rivals for each of those spots.13

This is the core idea behind answer engine optimization. Semrush's advice is to track brand mentions inside AI Overviews and to make sure both your own pages and third-party pages describe products accurately.12 Research from 2026 suggests brands cited in AI Overviews get 35% more organic clicks and 91% more paid clicks than competitors who aren't cited.10 Seer's numbers point the same way: pages cited in an Overview got 2.1% click-through, compared with 0.9% for pages that weren't cited.7

What it means

The common view in 2024 and 2025 was that AI Overviews mostly hurt informational content, while money keywords stayed relatively safe. The Semrush panel undercuts that view, even with its small sample and missing click data. Google appears willing to put AI summaries above organic results for the searches closest to a purchase, and especially for the most-searched ones.12

Semrush's claim that a number-one ranking "likely won't lead to as much traffic" as a year ago is not supported by its own data.13 Still, the independent click-through research makes it a reasonable conclusion.27 Brands will likely adjust by competing to be cited in AI answers. Independent review and affiliate publishers face a harder problem: Google is now summarizing the comparisons they used to write, often using their own work as source material. For them, this is a direct threat to the business model.