We've spent the last few weeks covering the agentic-commerce gold rush — Dick's coach-style chatbot, PayPal and Hey Savi's UK checkout agent, Salesforce's Agentforce release, Amazon renting out its AI assistant. So it's worth pausing on a new data point that complicates the story: most shoppers don't actually want AI to pick their clothes.
The discovery number is brutal
In a YouGov survey of U.S. clothing shoppers, just 6% said they'd be interested in using general AI tools like ChatGPT or Gemini to discover new clothing or brands. Asked how they actually prefer to find new styles, 60% said browsing in stores, 46% retailer websites or apps, 40% recommendations from friends or family, and 37% search engines. The chatbot finishes near the bottom of a list that's topped by the oldest discovery mechanism there is: wandering around a store.
YouGov, which polled about 957 clothes shoppers in late May, found the resistance isn't just unfamiliarity. Shoppers have practical doubts: whether the AI will actually work well, whether it will protect their data, and whether it will make shopping feel less personal. Those are not problems a slicker model fixes overnight.
Where AI is welcome: the boring stuff
The same survey points to a much more constructive read. The most-wanted AI use case was checking product availability or stock (26%), followed closely by size and fit recommendations (25%), with discovering products based on stated preferences trailing at 21%. Notice the pattern: shoppers want AI to remove friction from decisions they've already made — is it in stock, will it fit — not to make taste-level decisions for them.
That distinction is the whole ballgame for retailers allocating budget. An AI that tells a shopper the sweater they're eyeing is available in their size at the store two miles away is solving a real, daily annoyance. An AI that insists on curating their entire wardrobe is solving a problem most people don't think they have.
The discovery layer is fragmenting anyway
There's a structural wrinkle underneath the attitudes. A new Athos Commerce report argues that AI and fragmented discovery are reshaping fashion e-commerce, with shoppers scattering across more touchpoints and arriving with higher expectations. And it's not that consumers reject AI agents outright — a separate Chain Store Age survey found people are still using AI shopping tools, just with lingering skepticism about trusting them. The mood isn't "no." It's "prove it."
The takeaway for retail tech
The agentic-commerce investments aren't wrong, but the framing matters. The fastest returns are likely to come from AI deployed as a utility — inventory visibility, fit, reordering, service — where consumer demand already exists and trust is easier to earn. The harder, slower play is AI-as-tastemaker in discovery, where a 6% interest rate means you're not riding a wave; you're trying to create one.
For the vendors selling discovery agents into apparel, that's not a death sentence, but it is a reality check. Lead with the chores shoppers want done, build trust there, and earn the right to recommend. Push AI into the one place shoppers say they'd rather browse a rack, and you'll spend a lot of money discovering what YouGov just told everyone for free.