Only 10 percent of US consumers currently begin their product discovery with AI, according to the AlixPartners 2026 Consumer Sentiment Index, a survey of more than 9,000 American shoppers. The finding cuts against the hype around AI shopping agents: adoption is real but still narrow, concentrated among younger and higher-income shoppers, and most consumers remain unwilling to hand the actual purchase to an AI agent.
Adoption varies sharply by generation and income
The study found that 14 percent of Gen Z and millennial consumers start product search with AI, well above the 10 percent overall figure. Income matters even more than age: nearly 70 percent of consumers earning $150,000 or more say shopping with AI is important to them, compared with 53 percent of those earning under $45,000. The pattern suggests AI shopping tools are currently a premium-channel phenomenon, used most by shoppers with more to spend and more interest in concierge-style assistance.
Despite the low starting point, the research predicts a forthcoming uptick in adoption, particularly for commoditized goods. In fashion, brand and relationship will still matter at the top of the market, but the report foresees a future where queries for replenishment basics, such as black leggings, true-to-size, under thirty dollars, are offloaded more readily to AI agents. That would create a bisected retail landscape: the top end of the market remains brand-led, with the biggest concern being AI-powered copycats, while the bottom becomes spec-and-price driven, and brands caught in the middle must navigate both forces.
Metadata is becoming the new shelf
The most actionable finding for retailers and suppliers concerns product data. The report argues that AI agents do not read imagery or persuasive copy; they read structured data, including attributes, availability, and price. If a product's data is thin or messy, the agent either cannot surface the product or describes it badly, and the brand goes invisible in the fastest-growing channel there is. For merchants, that elevates metadata quality from a technical detail to a competitive requirement.
Other research points in the same direction, though with different measurements. A Snipp survey found 41.6 percent of consumers now start product searches with an AI assistant, while research from PYMNTS put the share of US adults beginning retail product research with AI at 19 percent and found that 39 million consumers have stopped using a traditional search channel they once relied on. An Emplifi and Semrush report found 66 percent of consumers have used AI to research products or discover gift ideas, and NielsenIQ data shows 55 percent of global consumers report some level of AI influence on consumer packaged goods discovery. The differing figures reflect different questions, but the direction is consistent: AI is ascending as a discovery channel even if it has not yet become the default.
The impact on purchasing behavior is already measurable. Among consumers who used AI to research retail products, 83 percent report at least one change to their purchasing decision: 43 percent found a better price, 27 percent chose a different brand, and 26 percent bought a different product entirely, according to the PYMNTS research. The same study found that Amazon closed 60 percent of consumers' most recent AI-assisted purchases, showing that AI is changing product discovery faster than it is changing checkout. A separate trust study found that among the 43 percent of US online shoppers who used AI for product research in the past 90 days, 86 percent verified the AI's recommendation through another source before buying, which means AI recommendations are influential but not yet decisive on their own.
Retailers are responding with their own assistants
Large retailers are not waiting for the channel to mature. Target, Best Buy, Macy's, and The Home Depot have all launched internal AI assistants, insulating themselves from the risk of being disintermediated as consumers shift from search engines to conversational tools. The study warns that the better AI agents become at suggesting products, the more consumers will trust them for other parts of the buying process, and the greater the risk that brands get sidelined from the customer relationship entirely.
For suppliers and e-commerce merchants, the practical takeaways are clear: structured, complete product data is the entry ticket to AI-driven discovery; pricing transparency matters more as comparison agents become common; and the distinction between brand-led and price-driven segments will widen. The report's own framing is the most direct: if the data is thin or messy, the agent either cannot surface you or describes you badly, and you go invisible in the fastest-growing channel there is. Retailers that treat metadata as a first-class investment today are positioning for the adoption curve the research says is coming.
The same research carries a warning for brands that rely on differentiation through imagery and storytelling. As AI agents become the discovery layer, persuasive content matters less at the moment of consideration and structured facts matter more; a brand with beautiful creative but thin attribute data can be described by an agent in ways that flatten its positioning into spec sheets. The report's prediction of AI-powered copycats at the top of the market only sharpens the point: in an AI-mediated channel, the durable moats are data quality, price competitiveness, and the trust signals that verification research shows consumers still seek before buying.
