What happened
- On September 10, Ant International, Mastercard and Visa announced they had begun collaborating on a Know-Your-Agent interoperability framework. The idea is for card networks, digital wallets and marketplaces to recognize an AI agent shopping on someone’s behalf.
- It seeks to connect three protocols each company had already launched separately: Visa’s Trusted Agent Protocol, Mastercard’s Verifiable Intent and Ant’s Agentic Mobile Protocol.
- The work will take place on BuildFin.ai, a platform convened by the Monetary Authority of Singapore, under that regulator’s safeguards framework for agentic finance. It covers Singapore’s payment ecosystems.
- Each network keeps its own verification and approval processes.
Why it matters
- For years, e-commerce anti-fraud systems were tuned to block automated traffic. If a verified agent reaches checkout, blocking it becomes a lost sale. A Chilean online store that filters bots by default today will have to tell apart the ones that carry credentials.
- The announcement sets accountability as a common goal, but it doesn’t say who’s liable if the agent buys something the user didn’t want. That’s the answer that matters for chargebacks.
- No Latin American payment gateway or acquirer appears in the announcement. The rules of agentic commerce are starting to be written in Singapore, with three companies at the table.
The number
$3 trillion to $5 trillion. That’s the consumer commerce agents would orchestrate by 2030, according to the McKinsey projection the announcement cites. What was signed to accompany that figure is the start of a collaboration in a single country.
Context
In August, AI agents started completing purchases on people’s behalf, and each payment network responded with its own protocol. Three different systems force a store or an agent platform to integrate three times. The agreement targets that cost: a verification all networks recognize, without any of them giving up its own risk assessment.
What’s next
- The BuildFin.ai working group is open to other companies, with no dates.
- No timelines announced for a common standard or for its application outside Singapore.
Bottom line
On September 3, Anthropic closed Claude sessions stolen by malware, a theft that bypassed the password and the second factor. Verifying the agent says nothing about who controls the session using it.
Sources
- Ant International, Mastercard and Visa Initiate Collaboration on Know-Your-Agent Interoperability to Scale Agentic Commerce — Business Wire, September 10, 2026
- Visa, Mastercard and Ant International team up on ID checks for AI agents — The Next Web, September 10, 2026
Edited by Rodrigo Cornejo. How we select and verify the facts, in who writes.
