Retailers rush to AI shopping agents, experts warn of risks

Handing your shopping to an AI agent may save a few clicks, but it can also cost you the last chance to spot a bad price, a wrong item or a risky data handoff before checkout. Technology experts say agentic commerce is still too new for many consumers to trust with full autonomy, even as the biggest names in retail and payments race to build it into everyday shopping.

Matt Kropp, an AI expert at Boston Consulting Group, put the warning bluntly: “It isn't mainstream yet and it's pretty risky right now.” That risk is not abstract. Amazon’s Rufus can track product prices, alert customers when prices hit a set level and complete purchases on Amazon. Walmart’s Sparky can help shoppers find products, read customer reviews and place orders. American Express this week announced new protections for cardholders using specified AI agents, saying it will verify an agent’s identity at checkout and shield eligible customers from charges tied to AI agent error.

The pitch is convenience, but the tradeoff is control. Visa’s research found that roughly two-thirds of consumers would use AI shopping agents to save time or find better prices, and nearly nine in ten want transparency into how those agents make decisions. About half said they would stop using the tools if that control disappeared. Visa surveyed 3,700 online shoppers in the United States, Australia and New Zealand in a 15-minute online survey, a sign that the company sees consumer trust as the central hurdle, not the software alone.

Retailers are moving fast because they fear being cut out of the sale. Etsy, Target and Walmart have pushed merchandise onto external AI platforms such as Google’s Gemini and Microsoft’s Copilot, even as analysts warn that the retailer can lose direct customer relationships and data ownership when a third-party agent becomes the shopper’s front door. Retail Dive reported that AI-driven U.S. e-commerce traffic grew 758% year over year between November 1 and December 31, 2025, underscoring how quickly the channel is expanding.

The money chasing the trend is enormous. SAP has cited McKinsey projections that the U.S. B2C retail market could see up to $1 trillion in orchestrated revenue from agentic commerce by 2030, with global projections ranging from $3 trillion to $5 trillion. Visa unveiled Intelligent Commerce Connect on April 8, 2026, and said it was in pilot with partners including Aldar, AWS, Diddo, Highnote, Mesh, Payabli and Sumvin. Mastercard said it launched Mastercard Agent Pay last year and is collaborating with Google, OpenAI, Cloudflare, PayPal and Microsoft.

But the cautionary tale is already clear. One tech founder reportedly asked an AI agent to secure a speaking slot at the World Economic Forum in Davos, and the bot succeeded, but booked an opportunity for $30,000, more than he could afford. In agentic shopping, that mistake could be a price glitch, a privacy leak or a purchase made too quickly to undo.