Visa, Mastercard, and Ant International—a Chinese financial technology company—are collaborating on a new initiative called Know-Your-Agent. This framework is designed to monitor and regulate AI agents that make purchases on behalf of users. The system aims to identify these AI agents, confirm they have proper authorization, and track their activity using signals such as identity verification and transaction data. While each company will maintain its own rules for approving or rejecting transactions, the initiative seeks to create shared standards across the global payment industry. According to Ryan McInerney, CEO of Visa, many consumers currently use AI for research related to purchases, but they generally do not trust AI to handle actual payments. A Visa study revealed that only 23% of U.S. internet users trust generative AI to manage payments, indicating a widespread lack of confidence in the technology. The Know-Your-Agent framework includes three key components: linking each AI agent to a verified human operator, ensuring the agent behaves securely through a common certification process, and monitoring its activity for any suspicious behavior. This initiative does not replace existing technologies already used by these companies. Visa currently uses the Trusted Agent Protocol, Mastercard employs Verifiable Intent, and Ant International relies on the Agentic Mobile Protocol. Additionally, EMVCo, a global standards body for payment technologies, has proposed its own framework to verify the intent behind AI agent payments, ensuring that AI respects payment limits set by users. Despite these efforts, concerns remain about AI hallucinations—when an AI agent misinterprets or misunderstands instructions, potentially leading to unauthorized transactions. Jiang-Ming Yang, director of innovation at Ant International, points out that hallucinations are a common and ongoing issue with AI that has yet to be fully resolved. However, the economic potential of AI agents is considerable. According to McKinsey, AI could generate between $3,000 and $5,000 billion in global trade by 2030, highlighting the importance of developing safe and trustworthy systems for their use.