Singapore-based Ant International, Mastercard and Visa have begun work on a shared Know-Your-Agent framework intended to help payment networks and merchants identify AI agents and connect their activity to an authorized person or business.
The companies said on September 10 that the planned interoperability framework would cover card networks, digital wallets, agent platforms and marketplaces. Each network would retain its own verification and decision-making processes, while recognizing common trust signals. Reuters also reported the initiative.
The announcement describes a framework under development, not a technical standard already deployed across the three payment networks. The participants did not provide an implementation timetable, pilot volumes or a list of merchants and financial institutions that will test it.
Common checks for agent-led payments
The proposed framework focuses on three areas. Cross-network operator traceability would link an AI agent to a validated operator, cardholder, company or other organization. Shared certification requirements would assess agents against agreed security and behavioral conditions. Continuous transaction monitoring would combine identity and payment signals for ongoing assessment.
The companies said interoperability could reduce repeated identity checks and integration work when an agent moves between wallets, card networks and marketplaces. It could also give merchants and payment providers a clearer record of who authorized an agent and whether its behavior remains within approved limits.
No certification criteria or technical specification accompanied the announcement. Questions about liability, revocation, data sharing, dispute handling and how networks will recognize one another’s trust decisions remain open.
Three existing protocols provide a starting point
The work builds on separate systems developed by each participant. Ant International released its open-source Agentic Mobile Protocol in April. The company said the protocol supports transactions involving digital wallets and had begun pilots with partners including Mastercard and Visa.
Mastercard’s approach includes Verifiable Intent, which is designed to carry evidence that an AI agent is acting within a user’s instructions. The company has also expanded its agent-payment work to machine-to-machine transactions through Agent Pay for Machines.
Visa introduced its Trusted Agent Protocol to let approved agents pass identity and purchase-intent information to merchants. The protocol is intended to distinguish authorized shopping agents from malicious or unwanted automated traffic.
Singapore platform will host the collaboration
The three companies plan to work through BuildFin.ai, an industry platform convened by the Monetary Authority of Singapore. Their effort will build on the regulator’s Safeguards for Agentic Finance at Runtime, or SAFR, framework for managing risks as AI agents carry out financial tasks.
Ant International, Mastercard and Visa said the collaboration will examine common approaches to agent verification, accountability and risk management in Singapore’s payment ecosystem. The current scope is framed as exploration toward shared principles; the organizations have not said whether the result will become an open standard or how it would be governed.
The initiative brings a major wallet network and the two largest global card networks into the same effort around agent identity. Its practical reach will depend on the technical rules, participating institutions and live transaction tests that follow.
Hong Kong selects 36 agentic AI pilots for financial sandbox