Chargebacks911 Founder and CEO Monica Eaton recently published a new guest feature in Finextra examining one of the emerging challenges surrounding agentic commerce: who is responsible when an AI agent makes a purchase that the consumer did not intend to make.
The article, When AI makes the purchase, who owns the mistake? looks at the growing use of AI agents to make purchases on behalf of consumers. Eaton argues that authentication alone cannot resolve the disputes that may arise when an AI agent makes a legitimate transaction that nevertheless produces an unwanted outcome. As AI takes on more purchasing decisions, merchants may face new challenges in determining what a consumer authorized, what an AI agent decided independently, and who is responsible when those two things do not align.
“Delegated authority is not unlimited intent,” Eaton writes. She explains that giving an AI assistant permission to manage routine purchases does not necessarily mean a consumer has agreed to every decision the agent makes about timing, quantity, or price.
Eaton identifies the resulting evidence gap as a major concern for the payments industry. Merchants may be able to show that a transaction was authenticated, fulfilled, and delivered, but they may not have access to the consumer’s instructions or the AI agent’s decision-making history. She argues that the industry needs clearer rules for agentic disputes, along with evidence and resolution processes that reflect how these transactions actually occur. The goal, she concludes, should be to build dispute and resolution infrastructure alongside the transaction systems that are making agentic commerce possible.
