Autonomous AI agents demand new payment standards

Waller identified agentic commerce—the use of artificial intelligence agents to make autonomous purchases—as an emerging shift for financial infrastructure.

He differentiated between agent-assisted models, where buyers retain final payment control, and agent-delegated models, where autonomous software executes transactions under preset guardrails.

While consumer applications are advancing first, Waller stressed that business-to-business transactions present greater efficiency gains alongside higher exposure risks.

Autonomous purchasing will also require continuous machine-to-machine micropayments across diverse rails, including cards, ACH, wire, and instant payment systems.

Trust mechanisms represent the primary hurdle to scaling delegated agents.

Large language models already reduce false positives in sanctions screening and improve payment routing, but Waller warned of cyber asymmetries where attackers need only one opening.

Furthermore, conventional fraud systems calibrated to human behavior cannot adequately monitor automated agent transactions.

He emphasized the necessity of interoperable technical standards over closed ecosystems to preserve merchant access and establish transparent liability when unauthorized purchases occur.