On June 10, 2026, Visa and OpenAI announced a strategic collaboration that places Visa’s payment infrastructure directly inside AI agent workflows. The partnership is not a pilot or a framework document. It integrates Visa’s tokenization, real-time authorization, and fraud monitoring capabilities into OpenAI’s platform, enabling AI agents to initiate and complete purchases on behalf of users within clearly defined permission boundaries.
The Architecture of Agent-Initiated Payments
The technical implementation relies on tokenized Visa credentials rather than raw card numbers. When an AI agent initiates a transaction, it operates within user-defined controls: spending limits, merchant categories, and required approval thresholds. Visa’s network handles the underlying authorization and fraud scoring in real time, applying the same risk infrastructure that processes billions of human-initiated transactions daily.
Jack Forestell, Visa’s Chief Product and Strategy Officer, positioned the stakes explicitly: “AI will transform commerce more profoundly than the internet or mobile technology ever did. As AI agents become active participants in the economy, Visa’s focus is to ensure transactions are trusted, secure and seamless.”
Why OpenAI Needs Payment Rails
For OpenAI, the partnership solves a practical constraint. AI assistants that can research, recommend, and plan are valuable. AI assistants that can also execute purchases become indispensable. Marco Mahrus, Head of Partnerships for Commerce at OpenAI, acknowledged the shift: “Commerce is going to happen in many more ways than it does today, and agents will play an increasingly important role in helping people complete tasks that involve money.”
The collaboration extends beyond consumer transactions. Visa and OpenAI will explore enterprise applications including developer-focused experiences powered by Codex and automated conversational workflows. This suggests that B2B procurement, not just consumer shopping, sits within the partnership’s scope.
Visa Intelligent Commerce as Platform Strategy
The OpenAI deal is the highest-profile deployment of Visa Intelligent Commerce, but it is not the only one. Earlier in 2026, Visa partnered with Fiserv to deploy Trusted Agent Protocol across Fiserv’s merchant acceptance ecosystem, authenticating and processing agentic transactions at the point of sale. The pattern is consistent: Visa is positioning its network as the trust layer that sits between AI agents and merchants, regardless of which AI platform initiates the transaction.
This approach carries strategic logic. Rather than building its own AI capabilities, Visa extends its existing strengths (tokenization, fraud detection, global network reach) into a new transaction category. Every agentic transaction that flows over Visa rails generates the same network economics as a human-initiated one.
The Competitive Landscape Sharpens
Mastercard has pursued a parallel strategy, partnering with FIS to enable agent-initiated transactions across its network. Stripe launched its Agentic Commerce Suite with Google and Meta integrations. The convergence of these announcements within a six-month window suggests that agentic commerce infrastructure is consolidating rapidly around established payment networks rather than being disrupted by new entrants.
The implication for merchants is straightforward: businesses that instrument their product catalogs and payment acceptance for agent-initiated transactions will capture demand that others miss. The infrastructure layer, as of mid-2026, is no longer theoretical. It is live, tokenized, and scoring transactions in real time.
Source: Visa Press Release, June 10, 2026
Related coverage: FIS Launches First Agentic Commerce Platform for Banks | Adyen Is Building a Commerce Operating System
