AI Mode Surfaces Products 21.6% More Expensive Than Traditional Search

Cheapest no longer wins. A study by Productrise, running identical shopping queries simultaneously through Google's AI Mode and traditional search, found AI Mode consistently surfacing more expensive products — by 21.6% on matched items.

For anyone competing on price in ecommerce SEO, this is the most consequential finding of the month.

What the study found

Productrise tracked more than 2 million product listings across over 100,000 search responses between 9 and 31 August 2026, comparing AI Mode against traditional search for the same queries, on the same dates, across US and UK markets.

On matched products — the same item, same query, same day — AI Mode showed listings 21.6% more expensive on average.

Across all listings, the gap was wider still: a median price of $149 in AI Mode against $100 in traditional search, a 49% difference.

Where prices disagreed on matched products, which happened in 38.1% of cases, AI Mode was the pricier of the two 68.4% of the time, with a median gap of 22.2%. On the rarer occasions AI Mode was cheaper, the difference was slight — a median of 7.8%.

That asymmetry is the detail worth sitting with. It isn't noise scattered evenly around a mean. When the two surfaces disagree, AI Mode skews expensive, and it skews expensive by a much larger margin than it ever skews cheap.

What's actually happening

There's an easy misreading here worth heading off: this is not Google showing the same product at a higher price. It's AI Mode selecting different products, from different retailers, that happen to cost more.

The mechanism, on the study's reading, is that lowest price is simply less of a ranking factor in AI Mode. What appears to carry more weight is feed quality, richness of product data, reviews, and how well a listing fits the conversational context of the query.

That's a coherent story. A model answering "what's a good running shoe for flat feet" is matching against a described need, not executing a price sort. Rich structured data gives it more to match on. A bare listing with a low price and three attributes gives it almost nothing.

Why this matters more than it first looks

Price-leader positioning loses its main SEO advantage. A retailer whose entire organic strategy is being cheapest has been well served by traditional shopping surfaces, where price comparison is explicit and prominent. In AI Mode, that advantage substantially evaporates. Being cheapest is not a signal the model appears to weight heavily.

Feed quality becomes the competitive moat. This lines up with the Merchant Center AI reporting Google shipped this week, which explicitly flags missing product attributes and recommends adding search terms to descriptions. Google is telling retailers what the model needs, and this study is independent evidence of what happens when you don't provide it. The two findings corroborate each other.

Premium and mid-market brands have an opening. If you've historically lost visibility to cheaper competitors on price-sorted surfaces, AI Mode is a surface where quality signals, review depth, and detailed product data can win instead. That's a genuine strategic opportunity, not a consolation prize.

Your reporting is probably blending two different worlds. If AI Mode traffic and traditional shopping traffic land in the same analytics bucket, you're averaging two surfaces with materially different selection logic — and different average order values. Segment them before you draw conclusions about which products are performing.

What to do

Run your own version of the test before accepting the finding wholesale. Take twenty of your priority product queries, run each in AI Mode and in traditional search, and record which of your products appear in each and at what price point. It takes an afternoon and tells you where your catalogue actually sits.

Then work on the inputs the model appears to reward: complete attributes, genuine review volume, and product descriptions that read like answers to what a shopper is asking rather than manufacturer boilerplate.

Caveats

This is a single study covering a three-week window in August, across US and UK markets, by a company whose product serves ecommerce merchants — a commercial interest in the finding being interesting. The sample is large and the methodology is sound in its comparison design, but one study establishes a strong signal, not a settled fact. The finding is worth acting on because the recommended actions — better feeds, richer data, real reviews — are things you'd want regardless of whether the price effect holds up.

Sources

• Productrise: Google AI Mode prefers more expensive products
• Google Merchant Center Help: AI Performance Insights