American Express announced new services and protections for cardholders who use specified artificial intelligence agents to make purchases. The offerings verify the identity of an AI agent at the point of sale and shield eligible customers from charges tied to AI agent errors.

AI agents can organize a user's email and purchase items based on budget and style preferences. Outsourcing those decisions to AI exposes consumers to risks including communication errors, financial losses, and potential data breaches. According to market research firm Statista, approximately 25% of Americans aged 18 to 39 have tried using AI to research products or shop. "The service will protect eligible customers from charges related to AI agent error," American Express said in describing the new offering.

Major retailers have rolled out their own agentic tools. Amazon's AI assistant Rufus can track product prices, alert customers when they reach a set level, and complete purchases. Walmart has deployed a conversational AI agent named Sparky that can help consumers find products, provide customer reviews, and assist with ordering.

Matt Kropp, an AI expert with Boston Consulting Group, said the technology is not yet broadly trusted for autonomous spending. "It isn't mainstream yet and it's pretty risky right now, because there aren't enough guardrails in the system for people to feel comfortable with agents autonomously buying things for them. It could potentially go buy a car, but I wouldn't say, 'Here's my credit card.'" Kropp said.

Andrew Lee, founder of Tasklet, said the underlying capability is already in place. "Agentic AI today is fully capable of shopping for people and doing normal things consumers can do," Lee said. He added that problems can arise when a user prompt gives an AI agent conflicting instructions.

Lee also expressed reservations about handing financial control to the software. "The agents are fundamentally hard to trust. Personally, I am not super comfortable with that yet. I like to control where my money goes myself, and as a business, we don't recommend that." he said.

Bretton Auerbach, founder of a New York-based tech startup, said bad actors can lure AI agents into disclosing a consumer's personal information. "If you give an agent your credit card and say, 'Go to this website and buy me something online,' there are ways to trick the agent," Auerbach said. "It might mistake a legitimate website for a phishing website that says in big, bold, text, 'Paste your credit card number here.'"