Visa has completed hundreds of secure, agent-initiated transactions in production and predicts millions of consumers will use AI agents for purchases by the 2026 holiday season. The payments giant is building infrastructure that allows AI agents — not humans — to authenticate, authorize, and complete purchases within defined spending rules.
The Trusted Agent Protocol
Visa developed an open verification framework called the Trusted Agent Protocol. AI agents register public keys in a Visa-managed directory. When an agent initiates a transaction, it cryptographically signs the HTTP request. Merchants and Visa verify the signature against registered keys to confirm the agent is authorized to act on behalf of a specific user with specific spending limits.
The system addresses the fundamental trust problem of AI commerce: how does a merchant know the AI agent placing an order is actually authorized by the cardholder? Visa’s answer is cryptographic identity at the agent level, with Akamai providing additional identity, authentication, and fraud controls.
Who’s Building
Over 100 partners worldwide are integrating with Visa’s agentic commerce infrastructure. Thirty-plus are building in Visa’s VIC sandbox. Twenty-plus agents and enablers are integrating directly. Closed-beta partners include Skyfire, Nekuda, PayOS, and Ramp.
Combined with Shopify’s Agentic Storefronts announced the same week, the entire commerce stack is being rebuilt for a future where AI agents are the buyers. Merchants will need to market to algorithms, not people. Payment processors will need to authenticate software, not humans. Fraud detection will need to distinguish between authorized agent transactions and unauthorized ones — a challenge fundamentally different from detecting stolen credit cards.
