Productrise tracked more than 2 million product listings over 23 days in August and found that when an identical product surfaced in both Google AI Mode and traditional search on the same day, the price shown in AI Mode averaged 21.6% higher.

The study was published on September 1, 2026 by Hugo Huijer, founder of Productrise, a platform that tracks organic product visibility inside Google. Over 23 days, from August 9 to August 31, 2026, the company ran product-centric shopping queries through both surfaces on the same calendar day and logged what each returned. The dataset covers more than 100,000 results pages and AI Mode responses in the United States and the United Kingdom.

One number carries the study. Where the same product appeared on both sides for the same query on the same day, the lead price in AI Mode was 21.6% higher on average than the lead price in traditional search. That is not a comparison of different product mixes. It is the same item, priced twice.

How the gap was calculated

The methodology is narrow by design. On the traditional search side, analysis was limited to listings inside the popular_products carousel, the free product block that sits in standard results. On the AI Mode side, Productrise used all tracked product cards. Products were matched across the two surfaces using Google's stable product identifier where present, restricted to the same query on the same calendar day.

For each matched pair, the company took the lead offer on each side, defined as the listing in the lowest position, then measured the AI Mode price against the traditional search price: the AI Mode price minus the results page price, divided by the results page price. United States prices are reported in dollars and United Kingdom prices in pounds, with no currency conversion applied.

That construction matters for reading the headline. The 21.6% figure describes the first offer a shopper sees attached to a product, not the cheapest offer available for it. According to Productrise, clicking a product on either surface opens a sidebar panel listing other sellers and their prices, some of them cheaper. The study's argument is about what happens before that click, which most shoppers do not make.

A higher-priced shelf, not just higher prices

Matched pairs are the cleanest comparison, but they are a small slice of the data. Productrise ran a second, broader analysis across every priced listing on each surface, including products that appeared on only one of them.

On that basis the typical AI Mode listing sits well above the typical traditional search listing. The median AI Mode product was priced at $149 against $100 in traditional search, roughly 49% higher, according to the study. The distribution chart shows the AI Mode side as visibly smaller, for a structural reason: AI Mode shows far fewer products. In the current dataset it accounted for 12.3% of all products tracked, an average of 3.9 products per response against 27.8 in traditional search.

Two effects therefore stack. Fewer items reach the shopper, and the items that do reach them skew expensive.

Overlap of 1.28%

The two surfaces respond to the same queries. They rarely respond with the same products. Across the study window, only 1.28% of products ranking in traditional search also appeared in AI Mode for the same search on the same day. Productrise calculated a match rate for each day, dividing matched products by all traditional search listings carrying a product identifier, then averaged the daily rates.

Expressed in absolute terms, the average query that produced results on both surfaces returned 27.8 traditional search products, 3.9 AI Mode products, and 0.94 products in common.

That figure sits alongside earlier measurement of the same surface. AI Mode returned roughly 95% fewer product listings than standard search in a Productrise study covering 21 days in July 2026, with standard search returning products on about 88% of tracked queries against roughly 23% for AI Mode, and product overlap reported at 0.8%. The new release describes that July work as having found AI Mode ranking about 5% of the products that rank in traditional search overall, a volume ratio rather than an overlap rate. The two studies use separate datasets and different denominators, so the 0.8% and 1.28% overlap figures are close but not directly interchangeable.

When the two prices disagree

Matched products do not always carry a price gap. When the same product appears on both surfaces, the two prices actually differ 38.1% of the time, according to the study. Where they differ, AI Mode is the more expensive side in 68.4% of cases, or roughly two-thirds.

The size of the gap depends on direction. When AI Mode is the pricier surface, the median difference is 22.2% and the average is 88.5%. Productrise attributes the distance between those two numbers to extreme outliers, giving the example of a used item in traditional search matched against a new one in AI Mode. When AI Mode is the cheaper surface, the median difference is 7.8% and the average 11.9%, a materially tighter spread.

Asymmetry is the finding here. The upside case, from a shopper's perspective, is small and bounded. The downside case is larger and occurs more often.

One figure in the release does not reconcile cleanly. Alongside the 38.1% disagreement rate, the study states that prices agree about 68% of the time, which would imply disagreement of about 32%. The complement of 38.1% is 61.9%. Productrise did not explain the discrepancy in the published material.

The seller changes on half of matched products

Price is not the only thing that moves between the two surfaces. On matched products, the main seller was different 49.6% of the time.

That mechanic is separate from re-ranking. A different lead seller means AI Mode selected a different offer for the same product, not a different position for the same offer. Price changes and seller changes often travel together, according to the study, which also documented matched pairs where the price was identical but the seller and the product title were not, and pairs where an offer ranked eighth in traditional search and first in AI Mode.

Callum Lockwood, Director of Organic Search at Re:signal, treated the swap as the more consequential result, saying the system appears to be picking a different one on its own criteria. Feed data and availability are the obvious inputs, he said, while noting that surfaces of this kind typically also weigh third-party citations, reviews and press beyond a merchant's own feed.

How shoppers arrive in AI Mode

The study includes a section on entry points, because the price a shopper sees first only matters if the shopper does not realise the surface has changed. Google places an AI Mode control on the standard search box, which opens the conversational surface while carrying the query across unchanged. Chrome on desktop exposes an AI Mode button in the address bar. Results pages carry an AI mode tab that leads into the same query on the other surface.

PPC Land has tracked those placements as they appeared. Google began testing an AI Mode button inside the homepage search bar for Search Labs users in June 2025. Access through Chrome's address bar, known internally as the omnibox, followed in September 2025, and a researcher documented in July 2026 that the AI Mode pill in the Chrome 147 omnibox routes every query to Google's hosted models rather than to the on-device Gemini Nano model shipped with the browser.

Scale gives those entry points weight. Google reported that AI Mode had passed one billion monthly active users in May 2026, with query volumes more than doubling every quarter since the United States launch and average query length running about three times that of a conventional search.

Practitioners read the data two ways

The release carries commentary from six named search practitioners, and the split in their readings is instructive.

Brodie Clark, an independent SEO consultant, questioned how far the top-level comparison travels, noting that the higher price may be the one displayed inside the AI Mode grid result while, in his experience, the retailer with the lowest price gets the click in the end. He framed the value of the research as consumer-facing: it indicates that products recommended in AI Mode may not weight price as heavily as practitioners assumed.

Katelyn Geary, Senior SEO Strategist at Break The Web, took the harder line, arguing that open market comparison is replaced by a curated path to higher-priced stock when an algorithm favours the more expensive item two-thirds of the time and substitutes the seller as well.

Jamie D'Alessandro, SEO Manager at JAKALA, read the same numbers as an opportunity, calling the pattern a small win for merchants who don't always want to compete solely on price, and observing from the screenshots that AI Mode appeared more likely to link to a brand's own website than to a third-party marketplace.

Tristan James, who works in SEO at EE, connected the finding to household budgets, noting that between 30% and 75% of consumers cite price as the most important purchase factor depending on the market, and concluding that 'Fact-checking' AI in shopping experiences just got a whole lot more important.

Jeff Collins, an SEO consultant, framed the compression as intentional commercial design, since Google has a clear incentive to move people from a question to a purchase faster and four products in an answer does that better than twenty.

Huijer's own reading treats the price signal as a redistribution of competitive advantage. If the model is not gating visibility on lowest price, he argued, brands that cannot win the race to the bottom may now have a better shot, provided their feeds, product data and reviews are strong enough for the system to recommend them with confidence.

Productrise sells tracking products for exactly the visibility problem the study describes, which is worth holding in view when reading that conclusion. Google was not quoted in the release and has not published a response to the findings.

Where this sits against Google's own pricing rules

The study measures the offer AI Mode selects, not a price set for an individual shopper. That distinction has become sensitive. When Google introduced shopping ads inside AI Mode on February 11, 2026, the surface already counted more than 75 million daily active users, and the company answered criticism of its commerce stack by pointing to merchant policies that prohibit displaying prices higher than those on a merchant's own website.

That criticism had arrived weeks earlier. Google's shopping AI drew accusations of enabling personalised upselling in January 2026, following the launch of the Universal Commerce Protocol, a characterisation the company disputed. Regulatory attention has since firmed up: the Federal Trade Commission approved a proposed enforcement policy statement on undisclosed personalised pricing on August 19, 2026, opening a 30-day comment period, after the Senate held its first hearing on surveillance pricing on August 4, 2026.

Nothing in the Productrise data speaks to personalisation. What it describes is selection: which offer, from which seller, at which price, gets promoted into a compressed answer.

Why it matters for the marketing community

The commercial consequence lands on three groups at once.

For retailers running Google Shopping inventory, an overlap rate of 1.28% means conventional carousel visibility carries close to no read-through into the conversational surface. That gap has been visible in the plumbing for months. Google connected eight Merchant Center attributes explicitly to conversational experiences, several carrying documentation notes naming AI Mode, and later added a Merchant Center report tracking brand visibility inside AI Mode with an attribute completeness score. Both point the same way: the selection layer reads structured data, and a sparse feed constrains it.

For merchants whose strategy has been built on being the cheapest listing, the 49.6% seller mismatch is the number to sit with. Winning the price comparison in a carousel does not carry over if the conversational surface picks a different offer on criteria that are not published.

For agencies and in-house teams, measurement is the immediate problem. Free listing reporting shifted when Google Merchant Center changed how organic traffic is reported on August 24, altering the baseline against which visibility loss can be judged. The paid and organic boundary has also thinned, with sponsored tiles entering the free listing grid in the Shopping tab in April 2026 and sponsored listings inside AI Mode becoming visually similar to organic recommendationsat a scale of a billion users.

There is a consumer-facing dimension that advertisers cannot control but will feel. AI Overviews and AI Mode both compress the comparison step that shoppers used to perform across multiple tabs. If that compression systematically surfaces the pricier offer, the trust premium attached to an AI recommendation erodes, and brands carry the reputational cost of a selection they did not make.

The narrower question for anyone managing product data is whether the criteria driving selection can be influenced at all. Productrise measured the output. The inputs remain undocumented.

Timeline

Summary

Who: Productrise, a platform tracking organic product visibility in Google, conducted the study; Hugo Huijer, the company's founder, published it. Commentary came from Brodie Clark, an independent SEO consultant, Callum Lockwood of Re:signal, Katelyn Geary of Break The Web, Jamie D'Alessandro of JAKALA, Tristan James of EE and Jeff Collins, an SEO consultant. Google, which operates both surfaces, was not quoted.

What: For products appearing in both Google AI Mode and traditional search on the same query and day, the lead price in AI Mode averaged 21.6% higher. Across all priced listings, the median AI Mode product cost $149 against $100 in traditional search. Prices differed on 38.1% of matched products, with AI Mode the pricier side in 68.4% of those cases, by a median of 22.2%. The main seller differed on 49.6% of matched products, and only 1.28% of products ranking in traditional search also appeared in AI Mode, which showed 3.9 products per response against 27.8.

When: Data was collected over 23 days, from August 9 to August 31, 2026. The study was published on September 1, 2026.

Where: The United States and the United Kingdom, the two markets Productrise tracked. Prices were reported in dollars and pounds without conversion.

Why: Google routes shoppers into AI Mode from the standard search box, the Chrome address bar and a tab on the results page, carrying the query across without a visible change of surface. If the first price shown there is systematically higher, shoppers can pay more without seeing the cheaper offer that standard search would have surfaced, and retailers competing on price lose the mechanism that previously earned them visibility.