Google’s AI Mode may be changing more than the format of shopping discovery. A new analysis from Productrise suggests it can also change the price point shoppers see.
Across 23 days in August, Productrise says it compared more than 2 million product listings from over 100,000 traditional Google search results pages and AI Mode responses, running the same shopping queries simultaneously. For products that appeared in both experiences, the lead price in AI Mode was 21.6% higher on average.
The study is an observational analysis rather than an independent audit of Google’s systems, and product matching, condition, seller selection and inventory changes can complicate price comparisons. Still, its scale and results make it a useful signal for merchants that treat AI search as an emerging storefront.
AI Mode appears to offer a narrower catalog
The sharpest finding may be how little overlap there is between the two interfaces. Just 1.28% of products shown in traditional search also appeared in AI Mode for the same query on the same day, according to Productrise.
Traditional search displayed an average of 27.8 products per query, compared with 3.9 in AI Mode. The overlap averaged fewer than one product per query.
That means AI Mode should not be understood simply as a conversational layer over the familiar shopping results page. It appears to be making a substantially different selection of products and sellers. On the study’s broader, unmatched comparison, the median AI Mode listing was priced at $149, versus $100 in traditional search—a 49% difference.
For shoppers, a smaller selection can reduce research effort, but it also leaves fewer visible alternatives. A buyer relying on the AI response may not encounter lower-priced offers that remain readily available in conventional results.
The seller often changes, too
Among matched products, Productrise found that prices differed 38.1% of the time. When they did, AI Mode displayed the higher price in 68.4% of cases. The median premium in those cases was 22.2%.
The study also found that the lead seller differed for 49.6% of matched products. That detail matters because a “same product” comparison can still represent different fulfillment, return, condition, availability or merchant-quality propositions. Productrise notes that extreme differences can include cases such as used inventory in traditional search versus new inventory in AI Mode.
In other words, the results do not establish that AI Mode is systematically selecting an inferior offer. They do show that price-conscious merchandising signals may not carry over predictably from traditional search to Google’s AI interface.
What retailers should do now
For merchants, the immediate implication is not to optimize around a presumed AI Mode preference for higher prices. The small product overlap indicates that ranking dynamics—and possibly the underlying retrieval and presentation logic—are still distinct enough to require separate monitoring.
Operators should consider:
- **Tracking priority queries in both experiences.** Record which SKUs, sellers, prices and product conditions appear, rather than assuming ordinary search visibility translates to AI Mode.
- **Auditing product-feed quality.** Keep titles, identifiers, images, availability, condition, shipping and price data consistent across first-party pages and feeds. In a curated result set, incomplete data can be costly.
- **Separating price from value signals.** Competitive pricing remains important, but fulfillment speed, returns, product detail and brand trust may influence which offer is selected or presented.
- **Checking channel consistency.** Promotions, marketplace listings and direct-store offers can create gaps that consumers interpret as a brand pricing problem, even when the offers differ legitimately.
What to watch next
The central unanswered question is why AI Mode’s results skewed pricier in this sample. Google could be weighing product relevance, merchant attributes, condition, availability or other signals differently from a traditional shopping surface. The observed gaps could also shift as AI Mode, product feeds and merchant participation evolve.
For now, businesses should treat AI shopping as a distinct discovery channel with its own measurement requirements. The risk is not merely losing rank. It is losing the chance to be considered when an AI interface reduces dozens of possible offers to a handful.
