Visa is positioning itself as the central nervous system for the nascent agentic commerce economy. By formalizing the Visa Intelligent Commerce platform, the world’s largest payment network is making a calculated infrastructure bet. If AI agents are to move from research assistants to autonomous purchasing entities, they require a standardized layer of trust, control, and connectivity that currently does not exist.
The platform’s architecture relies on a suite of tools designed to bridge the gap between machine-initiated intent and financial settlement. The Agent Score provides merchants with a metric to evaluate their website’s readiness for agentic interaction. The Agentic Directory serves as a verification layer, attempting to solve the mutual distrust problem by vetting both agents and merchants.
Visa CEO Ryan McInerney has been explicit about the primary hurdle: “As we introduce these new experiences, we’ve got to be equally focused on what do they reach for… for that sort of feeling of safety and trust so that they do want to engage in those new experiences.”
However, the gap between Visa’s infrastructure ambitions and current consumer behavior remains stark. Only 14% of consumers who used AI for product research trusted those recommendations without manual verification. The discrepancy highlights the central tension: Visa is building a sophisticated infrastructure, yet there is no public audit confirming what fraction of agent-initiated transactions represent actual buyer-seller commerce versus simple API plumbing.
Visa is betting that by providing the plumbing — the scores, the directories, and the tokenized security — it can manufacture the trust necessary to unlock a new category of volume. It is a speculative capital expenditure, deploying infrastructure in anticipation of a market that has yet to materialize.