Agent identity
The verification mechanism proving that an AI agent claiming to act on a shopper's behalf actually has that authorization. Managed via TAP + AP2 + upstream trust vendors.
Agent identity is the trust primitive underneath all of agentic commerce. When a merchant receives a checkout request from an agent, they need to know: is this agent genuinely acting for the shopper it claims to represent? Or is it a fraudster spoofing an agent header?
Agent identity is verified through a combination of TAP (Trusted Agent Protocol) signatures, AP2 payment-authorization tokens, and upstream trust providers (HUMAN, Signifyd, Riskified, Forter). Each layer catches different fraud vectors.
For most merchants, agent identity is invisible plumbing. Their agentic commerce platform handles verification; they just see clean, authenticated transactions. But brands with regulatory exposure (age-restricted products, controlled substances) or custom fraud logic need to understand the identity layer directly.
See also
TAP (Trusted Agent Protocol)
Protocol for verifying agent identity + authorization scope in commerce transactions. The trust-and-verification layer for agentic commerce.
AP2 (Agent Payments Protocol)
A cross-industry protocol standardizing how AI agents authenticate, authorize, and settle payments on behalf of shoppers. Payments-layer counterpart to ACP and UCP.
Scoped payment token
A single-transaction payment authorization used by an AI agent - limited to a specific merchant, amount, and time window. The primary payment primitive for agentic commerce.