72% shop with AI. 53% rebuy from last year's stores.

Optimove's Holiday Shopping Report 2026 landed with a headline number every AI-commerce vendor is about to put on a slide: 72% of surveyed U.S. consumers consult tools like ChatGPT, Claude, or Gemini for shopping ideas. The number two rows down is the one that should actually change your plan — 53% say they'll shop only at retailers they bought from last year. AI-powered discovery is real. It is not, on this evidence, doing much prospecting for you.

What actually happened

Optimove Insights published the report on Aug 5. Read the methodology before the findings: 648 U.S. consumers, 18+, household income $75,000 or more, surveyed in Spring 2026. That's a small, affluent, U.S.-only sample. Directionally useful, not a census.

The AI numbers break down further than the headline. Of that 72%, 49% use AI regularly and 23% occasionally. On specific features: 55% have used product recommendations, 41% chatbots, 31% visual search, 26% personalized deals, 23% size and fit. Roughly three-quarters trust AI-generated recommendations as much as or more than other sources — and 37% name data privacy as the thing that would push them to use AI assistants more.

Now the counterweight. 53% plan to shop exclusively at last year's retailers; 47% are open to new brands. Amazon (81%), Walmart (71%), and Target (58%) dominate stated preference. Quality (81%) and price (70%) drive purchases generally, and both jump for new brands — 85% and 80%. Spending is holding up: 55% expect to spend more than last year, 30% the same.

Why AI-powered discovery matters for your business

Put the two numbers side by side and the strategy writes itself. Shoppers are asking an AI, and then buying where they already bought. Which means AI discovery is mostly working as a retention and re-entry channel, not an acquisition one — an assistant is helping someone find the thing at the store they already trust.

That reframes where the money goes.

Your existing customer list is the AI channel. If half your buyers won't leave last year's retailers, and you're one of them, your job is to be findable and answerable when the assistant gets asked. That's clean, structured product data with real specs, stock, and returns policy — the same catalog work that converts. It's unglamorous and it's the whole game.

Quality and price go up as barriers, not down. For new brands, quality matters to 85% and price to 80% — higher than for incumbents. An AI assistant will surface your competitor's spec sheet next to yours in the same breath. If your product page can't defend the number, the assistant does the comparison for the shopper and you lose it silently.

37% naming privacy is a design input. That's the largest single lever on adoption in this sample. If you're bolting a shopping assistant onto your storefront, say plainly what it stores and what it doesn't. Labeling and disclosure are already law in the EU; here they're also conversion.

One caution we'd apply to our own reading: 648 affluent U.S. respondents is thin, and self-reported AI use runs ahead of measured behavior. Other data has shown AI-referred checkout conversion still under 1%. Fund the catalog work, which pays off either way. Don't fund a replatform on a survey.

Sources: Optimove Insights via GlobeNewswire, Optimove Holiday Shopping Report 2026.