24% of shoppers now compare prices with AI, not Google

A quarter of American shoppers are already asking a chatbot what your product should cost. Akeneo published survey results on September 23 from 1,000 US consumers polled by Dynata in August: 24% already use tools like ChatGPT or Google Gemini to compare prices, and 56% trust those tools to give accurate pricing. Another 24% expect to use them through the holiday season. AI price comparison stopped being a 2027 problem sometime this summer.

What actually happened

The trust numbers are the spine of this survey. Only 32% of consumers completely or mostly trust retailers to offer a fair or competitive price. Against that, 56% trust an AI tool's pricing answer. Shoppers extend more credit to a model that scraped your site than to your site.

The rest of the data describes people who verify everything:

• 77% have spotted the same product at different prices across retailers or platforms
• 79% have delayed a purchase waiting for a price to drop
• 59% say price matters more than it did six months ago
• 46% compare prices across multiple retailers when shopping online; only 9% buy without comparing
• 68% at least sometimes check a website or app for a lower online price while standing in your store

And the one that should shape policy, not marketing: 57% would trust a retailer significantly less if they learned prices had changed based on their personal information or shopping behavior. Akeneo CEO Romain Fouache's framing — pricing "can no longer sit in a silo from the rest of the product experience" — is a PIM vendor's pitch, and it happens to be correct.

For the holidays, 40% plan to compare prices across retailer websites, 37% will use search engines, and 24% will ask an AI tool.

Why AI price comparison matters for your business

The model reads your structured data, not your design. When a shopper asks Gemini to compare your product against three competitors, the answer is assembled from whatever is machine-readable: your Product schema markup, your feed, your title and spec fields, your offers price and availability. If your price lives only in rendered JavaScript, or your title is "NEW! Best Seller! Premium Widget (Blue) - Free Ship", you have handed the comparison to the competitor whose data is clean. That is not an SEO nicety anymore; it is whether you appear in the answer at all.

Stale data is now a pricing lie. A cached price in a feed, an out-of-stock item still listed as available, a sale that ended Tuesday — the model will repeat all of it confidently to a shopper who trusts it more than they trust you. Feed freshness has become a trust surface. Audit how often your product feed actually updates versus how often your prices change.

Personalized pricing is a 57% liability. Vendors will sell you dynamic pricing keyed to browsing behavior this quarter. The survey says most of your customers would trust you less for doing it, and shoppers who compare across four sites and a chatbot are exactly the people who will notice. Price by segment, volume, channel or time if you like — not by what you know about the individual.

Assume the price objection arrives pre-loaded. With 68% price-checking inside your store and 46% comparing online, the question is no longer "what does this cost" but "why is yours $12 more." Make sure the answer is in your product data — warranty, lead time, support, bundled install — and not just in a salesperson's head. Same for your site: the differentiator needs to be a structured field, because that is what the model can read.

Sources: Akeneo via PR Newswire.