AI is becoming part of the shopping experience, and consumers already have opinions about where it belongs.
New research from Net Conversion found that 54% of consumers now use AI at least sometimes to research purchases. The same survey found that 54% trust a brand less when its advertising looks AI-made. Meanwhile, research from Bazaarvoice and Bluefish found that authentic reviews and ratings are the biggest factor in whether consumers trust an AI product recommendation.
The difference comes down to which side of the transaction the AI is on. Consumers seem perfectly comfortable bringing AI into the buying process when it helps them research products, compare options and narrow the field. Their enthusiasm drops when the AI is coming from the brand instead.
The customer, in other words, brought their own AI along for the shopping trip.
Consumer Trust in AI Shopping Depends on Verifiable Evidence
The Bazaarvoice finding says a lot about what consumers expect an AI recommendation to be based on. Authentic reviews and ratings ranked ahead of brand reputation, price and the AI’s own explanation for its recommendation. Consumers appear willing to let AI interpret the market for them, provided that interpretation is grounded in something beyond what the seller says about itself.
That makes reviews, ratings, product specifications and other verifiable information more important as AI becomes part of product discovery. An AI shopping assistant can pull those inputs into a recommendation without relying on the brand’s carefully constructed path to purchase. And consumers will let AI make the case — as long as it's working from more than the marketing.
AI Shopping Agents Are Becoming Part of Product Discovery
AI is also becoming a filter between shoppers and the brands trying to reach them. An AI shopping agent can research products, compare offers and narrow a field before the shopper visits a product page. In some cases, the consumer may never see most of the brands the agent considered.
Research from Harvard Business Review suggests traditional marketing techniques don’t necessarily transfer neatly into that environment. At the same time, research from Wharton’s Generative AI Labs found that shopping-agent recommendations can change based on reviews, competing information, ordering and how information is retrieved. Keep in mind, shopping agents can still be influenced. Reviews, product information and even the order in which they encounter information can affect what they recommend.
But brands have two audiences now: the person making the purchase and the software deciding what reaches them. Both can be persuaded — just not by the same things.
AI Advertising Faces a Different Consumer Trust Problem
Brands have spent the past few years debating how much AI should be involved in making marketing. Meanwhile, consumers have started using the same technology to decide whether any of it is worth buying.
The latest research suggests consumers care quite a bit about what the AI is doing and who it’s doing it for. Consumers appear more accepting of AI when it works on their behalf and draws from evidence they can evaluate. Suspicion rises when AI appears to be speaking on behalf of the seller.
That gives marketers another reason to pay attention to what exists beyond the ad itself. An AI shopping agent can check a product claim against specifications, reviews and what customers are saying elsewhere. All of that becomes part of the pitch — the reviews, the specs, the third-party coverage — most of which the marketer never touched.
That matters to RAD because so much of our work starts with understanding what actually influences an audience before a marketing decision gets made. Now there’s another layer to account for. The customer’s AI may be doing some of that evaluating first, using information the brand doesn’t control. That makes the judgment before the ad even more important. Marketers still need to understand what actually matters to the audience, because generating a persuasive message is getting a whole lot easier than knowing what’s worth saying.
AI Shopping Is Changing Who Controls the Customer Journey
For years, brands controlled much of the path between discovery and purchase. They built the website, bought the ad, arranged the product page and decided which message appeared first. Now the customer can skip a good amount of that journey.
Amazon’s recent decision to block Meta’s Muse shopping agent shows that platforms are already starting to draw lines around where shopping agents can go. Amazon said third-party agents should identify themselves, respect a platform's decision about whether to participate and give customers the ability to opt out. The standoff is ultimately a fight over who gets to sit between the customer and the seller.
For marketers, the implications arrive before fully autonomous shopping does. A customer doesn’t need to hand an AI agent their credit card for the agent to influence the sale. The moment it researches the category, compares the options or builds the shortlist, the part of shopping that decides the outcome has already happened offstage.
Brands can keep debating how much AI belongs in their marketing, but customers are already deciding how much AI belongs in their shopping. That leaves marketers with a stranger job than the one they trained for. They still have to earn the customer’s attention and trust, except now their case may have to survive an AI assistant doing a little research of its own.
Marketers can produce more ads, faster, in a thousand variations. An AI shopping agent can look at all of it and recommend something else. Brands that miss the shift may end up very good at making marketing for an interface their customers no longer use.
