From Fortune · Nicholas Gordon · · 1m 30s
Banks must learn to know their AI agents
As autonomous AI agents handle trillions in payments, financial firms face a compliance crisis: their systems weren't built for non-human actors.
In brief
As autonomous AI agents handle trillions in payments, financial firms face a compliance crisis: their systems weren't built for non-human actors. Banks need to rebuild their infrastructure and compliance frameworks to identify, authorize, and monitor AI agents the way they do human customers, or risk massive…
The next compliance frontier
Banks have mastered know your customer rules. Now they face a new challenge: identifying and authorizing AI agents that shop and pay autonomously on behalf of users.
“In the agent economy, you need to know your agents. Who's the agent? Who does it belong to? Who authorized it?”
Agents could control trillions
McKinsey estimates AI agents could orchestrate as much as five trillion dollars in global consumer spending by 2030, making their oversight a critical financial stability issue.
Infrastructure wasn't built for this

Payment systems were designed for predictable human behavior. AI agents are probabilistic and adaptive, meaning the same request can produce different results, breaking core banking assumptions.
“Payment rails rely on predictable rules, legal certainty, and clear accountability structures to ensure trust and financial stability.”
Industry launches verification framework
Ant International, Mastercard, and Visa announced a know your agent interoperability framework to help card networks, digital wallets, and marketplaces recognize trusted AI agents across different systems.
Scale of errors multiplies with autonomy
JPMorgan's Benson Wong warns that agent failures don't just cause embarrassment. When autonomous workflows go wrong, the impact scales dramatically across transactions and systems.
“If you don't get agentic workflows correct, there are vastly scaled impacts.”
Know agents better than customers
Ant Digital's Zhuoqun Bian concluded that financial institutions must understand their AI agents even more thoroughly than they understand their human customers to maintain control and stability.
“You need to know your agents much more than you know your customers.”
The short version
Banks need to rebuild their infrastructure and compliance frameworks to identify, authorize, and monitor AI agents the way they do human customers, or risk massive operational failures at scale.



