Agentic AI has a toehold in B2C transactions, but has a ways to go to capture B2B transactions, Waller said at Sibos in Miami.
Artificial intelligence and autonomous AI agents will fuel the next evolutionary leap in global and cross-border payments, Federal Reserve Board Gov. Christopher Waller said during a speech at Sibos 2026 in Miami.
The payment industry’s early adoption of machine learning and large language models has helped combat payment fraud and quicken reconciliation and similar tasks, said Waller.
“Now it is helping build the foundation infrastructure for AI agents to operate more broadly in the economy,” he added. “These agents can plan and execute multistep processes using LLMs, enabling those agents to transact autonomously.”
Waller separated agentic commerce into two models: agentic-assisted and agentic-delegated. In the agentic-assisted model, the agent is primarily for product search and discovery, while the buyer remains in control. In the agentic-delegated model, “a buyer grants authority to an AI agent to shop and make payments on their behalf. The buyer may specify some constraints and set up guardrails, but the agent operates autonomously.”
Because agentic-delegated commerce could increase the risk of unintended purchases, Waller suggested deploying more extensive trust mechanisms and guardrails.
Agentic B2B Payments
Although agentic commerce has gained traction in consumer-to-business transactions with personal AI assistants, Waller said that business-to-business purchases would be fertile ground for agentic-delegated commerce.
“The purchases are often recurring and abide by a set of rules, such as approved suppliers and budget limits, which create natural guidelines for agents,” he said. “Agents could also potentially negotiate terms with suppliers and develop payment strategies to optimize working capital.”
However, the industry would need to develop and deploy robust controls and monitoring capabilities since B2B transactions typically carry a higher value than their consumer counterparts.
Also, unlike consumer-to-business transactions, AI agents executing B2B transactions will need to support a broad range of payment rails, including the Automated Clearing House, wire transfers, instant payments, and credit cards.
The biggest barrier to agentic commerce adoption is building trust mechanisms between buyers and sellers, which changes the payment authentication paradigm, said Waller. “The question shifts from proving that a buyer is an authorized payer to proving that an agent has the authority to pay on the buyer’s behalf. Capturing this will require new authentication approaches.”
He also noted two key questions about the future of agentic commerce: whether it will migrate toward platform-specific or interoperable standards. The latter would work across ecommerce systems, agent interfaces, and payment methods. “In this sense, they could help to level the playing field for smaller merchants and payment providers seeking to operate in this space.”
The payment industry still needs to take significant steps before it can bring agentic commerce to fruition, including developing standards for how agents can “carry identity, consent, and payment credentials across the full ecommerce stack,” as well as striking the proper balance between convenience and friction regarding agent authorization, Waller said.
These steps will require the payments industry to balance innovation with the safety, integrity, and stability that underpin trust in payments, he concluded.
Leave a Reply