Visa, Mastercard and Ant International draft agent ID standard

Visa, Mastercard and Ant International are building a shared standard for verifying AI agents before those agents can spend money on a person's behalf. The three companies announced the effort September 9-10, calling it know-your-agent, or KYA.

The problem they're solving is re-onboarding. Right now an agent verified to transact on one payment network has to prove itself again on the next network, the next wallet, the next marketplace. KYA interoperability would let one verification travel across all of them.

The companies are citing projections of $3 trillion to $5 trillion in agent-orchestrated consumer commerce by 2030. That's the number justifying the investment.

But Visa's own research undercuts the urgency. Only 23% of US consumers say they trust generative AI to handle payments for them. Building shared plumbing for agent commerce assumes people will hand agents their wallets. Right now, three out of four won't.

That gap is the real story here. A verification standard solves a coordination problem between Visa, Mastercard and Ant International. It does nothing to fix why most consumers don't want an agent buying things for them in the first place. Fraud liability, accidental purchases, and a lack of visibility into what an agent actually bought and why are all separate problems from proving an agent is who it says it is.

Watch whether any of the three publishes consumer-facing controls, like spend limits or purchase logs, alongside the identity work. A standard aimed only at network plumbing, with nothing addressing the 23% trust number, is a bet that trust will follow infrastructure. History with new payment rails doesn't support that bet.