Banks now need to ‘know your agent’, not just ‘know your customer’
Visa, Mastercard and Ant International are building a shared way to verify the AI agents that shop and pay on people's behalf.
When an AI agent buys something for you, the payment network needs to know three things. Which agent is it? Who does it belong to? Is it allowed to spend? In September, three of the world's largest payment players agreed to answer those questions together.
What happened
Visa, Mastercard and Ant International started work on a ‘Know-Your-Agent’ framework, convened through a Singapore central bank industry platform. It has three parts. Every agent is traced back to a verified owner, agents are certified against shared security standards, and transactions are monitored continuously. Each company's own agent protocol stays in place, but trust signals will work across networks.
Why it matters
Payment rails were built for humans who behave predictably. Agents don't. McKinsey projects that agents could orchestrate $3–5 trillion of global consumer commerce by 2030, so identity and liability for agents become a core compliance question.
What it means for your business
For banks and fintechs, this is a new KYC workstream covering agent onboarding, mandates, limits and dispute handling. For merchants, it's time to make checkout and product data readable by agents, not just by humans.
Sources: Fortune (Sep 19), Unite.AI, TNGlobal (Sep 10–11), PYMNTS