Shoppers who open Google’s AI Mode instead of a standard search see a higher starting price for the same product 21.6 percent of the time, on average, according to a Productrise study published September 1. The gap is not a mix of different items. Productrise matched identical products using the stable identifier Google assigns each listing, tracking more than 2 million listings across the United States and United Kingdom over 23 days, from August 9 to August 31.

The stakes go beyond a pricing quirk. AI Mode, Google’s conversational search experience, now routes over a billion monthly users through an entry point built into the standard search bar, a Chrome browser button, and a tab sitting inside the results page, carrying each query into a surface that shows roughly seven times fewer products than a standard results page.

Founder Hugo Huijer ran the comparison two ways. When the same product turned up on both surfaces on a given date, its lead AI Mode offer ran 21.6 percent above its traditional search counterpart. Across every priced listing, regardless of overlap, AI Mode’s median price reached $149, versus $100 for the same category in traditional search, a 49 percent gap driven partly by AI Mode simply showing fewer, pricier items: 3.9 products per response versus 27.8.

Overlap between the two surfaces is thin. Only 1.28 percent of the listings that rank in a standard search result also turned up inside AI Mode when the same query ran on the same day, close to the 0.8 percent Productrise measured in a July study using a different methodology. That earlier report found AI Mode returning roughly 95 percent fewer listings than standard search. The new data adds a second layer: fewer products reach shoppers, and the ones that do skew expensive.

When matched prices actually diverge, which happens 38.1 percent of the time, AI Mode comes out costlier 68.4 percent of the time, with a median premium of 22.2 percent. The reverse case, where AI Mode is cheaper, produces a tighter median gap of 7.8 percent. That asymmetry means the downside for shoppers who trust the first price shown is both more frequent and larger than the upside.

Seller identity moves too. On matched products, the lead seller differed between the two surfaces 49.6 percent of the time, a separate signal from price. Callum Lockwood, Director of Organic Search at Re

, called that swap the more consequential finding, since it shows AI Mode choosing offers on its own criteria rather than reordering the same sellers.

Practitioners split on what the pattern means. Katelyn Geary of Break The Web read it as replacing open price comparison with a curated, higher-priced path. Jamie D’Alessandro of JAKALA called it an opening for merchants who compete on factors besides price. Brodie Clark countered that shoppers who click through still tend to find the cheapest seller in the sidebar panel that follows, limiting how much the headline number matters at checkout.

The study measures selection, not personalization, and Productrise’s release did not include a Google statement. That distinction matters because regulators are already circling adjacent territory: the Federal Trade Commission approved a policy statement on undisclosed personalized pricing on August 19, and the Senate held its first surveillance-pricing hearing on August 4. Productrise’s data does not touch personalization, but it lands in a climate where any systematic price pattern in an AI surface draws scrutiny.

For merchants who have built visibility strategies around being the cheapest listed price, the figure to watch is that sellers differ 49.6 percent of the time. A strong Merchant Center feed with complete attributes, reviews, and third-party citations now competes for a different kind of selection than winning the standard carousel, and teams should audit their AI Mode presence separately from traditional Shopping performance rather than assuming carousel rank predicts it.

PPC Land, citing Productrise’s September 1, 2026 study, first reported these findings on September 3, 2026.