Will agentic commerce platforms adopt open or closed models?

Christopher J. Waller, in a Federal Reserve Board speech, proposes that agentic commerce platforms will split along the same open/closed axis as payment networks: "a related consideration revolves around open versus closed agentic commerce systems, which share similarities with open- and closed-loop payment networks." He defines the terms concretely: "An open system in this context is one that allows a range of shopping agents on its platform. A closed system is one that requires using a specific AI agent. For example, an online retailer might require shoppers to use its proprietary AI agent and block other agents."

Waller ties the choice to opposing structural consequences: "Open systems support buyer choice and an active environment for developing new agentic commerce solutions. Closed systems give online retailers more control over what happens on their platforms and lead to more vertical integration." He links this to a parallel standards question — "whether agentic commerce will migrate toward platform-specific or interoperable standards" — noting that "platform-specific standards are designed around a single ecosystem" while "interoperable standards are designed to work across e-commerce systems, agent interfaces, and payment methods," which "could help to level the playing field for smaller merchants and payment providers seeking to operate in this space."

This open/closed framing names the structural choice behind Will brands succeed by building their own AI shopping agents?: Accenture's advice — become the agent of choice or be chosen by one — is the brand-side strategy response to exactly the closed-versus-open platform split Waller describes, since a brand running a closed, proprietary agent is betting on vertical control while one accepting third-party agents is betting on being discovered by them. It is also the market-structure counterpart to Why do consumers delay adopting agentic search?: a platform that finds agentic search revenue-superior for itself has a concrete incentive to run a closed system, since closed systems are what "give online retailers more control" over that revenue relationship. And Can a protocol split agent access from agent incentives? is a working example on the open/interoperable side of Waller's divide, built to work across web, API, and agent channels rather than lock users into one proprietary agent.

Waller is explicit that this is unresolved framing, not a forecast: "at this early stage, it is not clear which model will have the edge in shaping agentic commerce, but whichever gains traction could significantly influence how the market structure evolves." The excerpt gives no data on current adoption of open versus closed agentic commerce systems, no named examples of which major retailers have chosen which model, and no mechanism for predicting the outcome — only the claim that the choice itself, however it resolves, will shape the resulting market structure.