Visa, Mastercard and Ant International have begun collaborating on a shared "Know Your Agent" (KYA) interoperability framework, an attempt to answer a question that agentic commerce is about to make very pressing: when an AI buys something on your behalf, which network vouches for it, and how does a merchant know it is acting on your intent?

Announced in São Paulo on 10 September, the initiative aligns three protocols the companies rolled out separately over the past year: Visa's Trusted Agent Protocol, Mastercard Verifiable Intent, and Ant International's Agentic Mobile Protocol. The goal is not to merge them but to make their trust signals readable across networks. In practice, an agent certified by one participant should not have to repeat the same identity and security checks to work with another — a problem that has kept agentic payments stuck in pilot programmes.

The framework rests on three commitments: cross-network operator traceability, so that every agent is linked to a validated operator, cardholder or business; shared certification requirements, so that agents are assessed against common security and behavioural criteria; and continuous transaction monitoring, using identity and transaction signals to keep that certification current.

The economics explains the urgency. The companies cite McKinsey projections that AI agents will orchestrate between $3 trillion and $5 trillion of global consumer commerce by 2030. Ant International's chief innovation officer, Jiang-Ming Yang, told CNBC that the potential for hallucination is precisely why trust has to come first: "we need to make sure people feel safe."

Visa and Mastercard dominate card payments in developed markets, but a large share of commerce in emerging economies runs through e-wallets — more than 50 of which connect to Ant International's Alipay+ gateway, reaching roughly 2 billion accounts. As those wallets increasingly link to cards, a common recognition layer matters more than any single company's protocol.

The work will continue through BuildFin.ai, an industry platform convened by the Monetary Authority of Singapore, building on the Safeguards for Agentic Finance at Runtime framework. What is still missing is exactly what a standard needs most: published technical specifications, a governance body and a rollout timetable. And trust is not only a technical problem — figures cited in coverage of the announcement suggest only about 14% of consumers currently trust AI to complete a purchase without verification.

Also worth noting: Alipay, the mainland Chinese payments app run by Ant Group, from which Ant International spun off, said this week that users can set recurring requests through its AI assistant — for example, ordering a Starbucks iced Americano every morning at 10am, with a prompt to confirm payment each time. The agentic future is arriving in small, quotidian increments.

Sources