Thursday , December 4, 2025

Google Launches Its Agent Payments Protocol to Validate AI Agents Are Making Purchases

Google Inc. early Tuesday launched its Agent Payments Protocol, an open protocol to authenticate consumers who are initiating a payment using an AI agent when shopping online.

Developed with 60 payments and technology companies, including Adyen NV, MasterCard Inc., and PayPal Holdings Inc., the protocol is intended to remove a key hurdle to agentic commerce—validating an AI agent has the authority to initiate a purchase on behalf of the shopper. Authentication of AI agents reduces the risk of these agents being used to commit fraud.

Agentic commerce is a new commerce model that enables an artificial-intelligence system to handle commercial transactions on behalf of users. Google’s Agent Payments Protocol (AP2) is an open, shared protocol that provides a common language for secure, compliant transactions between agents and merchants and that supports multiple payment options, including credit and debit, stablecoins, and real-time bank transfers.

To authenticate an AI agent, AP2 uses what Google calls Mandates, secure, cryptographically signed digital contracts that are verifiable proof a consumer has authorized the agent to act on her behalf. Mandates address the two primary ways a consumer will use an AI agent to shop: when the consumer is present during the transaction and when the consumer is not present.

A consumer-present transaction occurs when a shopper asks an AI agent to find her a new pair of white running shoes, for example. The shopper’s request is captured in an initial Intent Mandate that provides “auditable context for the entire interaction in a transaction process,” Google says. Once the agent presents a shopping cart with the shoes the shopper wants, she approves the purchase by signing a Cart Mandate “that creates a secure, unchangeable record of the exact items and price,” Google adds.

A shopper-not-present environment occurs when an AI agent is instructed to perform a task at a later date on the shopper’s behalf, such as buying concert tickets when they go on sale. In these instances, the shopper signs upfront a detailed Intent Mandate that specifies the rules of engagement, such as price limits and preferred seating locations. The Intent Mandate serves as pre-authorized proof that an AI agent is authorized to make the purchase once the preset conditions of the purchase are met.

“As agentic commerce becomes a significant part of the future of commerce and transactions, it introduces a new interaction model that requires a shared, open protocol, or common language, for proving authorization and authenticity,” A Google spokesperson tells Digital Transactions News by email.  

The use of mandates creates “foundational, auditable proof of user intent at each step. This proof helps mitigate fraud, enables financial institutions to better manage risk, and allows the entire ecosystem to innovate on new commerce experiences based on a foundation of trust,” the Google spokesperson adds.

AI agents from AP2 partners will be available in Google’s AI Agent Marketplace.

“For AI to truly drive commerce, agents need a secure and universal way to handle payments. Google’s new Agentic Payments Standard is a huge step forward, providing the foundational framework to make this possible,” Bam Azizi, chief executive and co-founder of Mesh, a San Francisco-based payments network for cryptocurrency and other digital assets, says in a statement.

Check Also

OneDine Signs California Pizza Kitchen for Payments Tech at 100-Plus Locations

California Pizza Kitchen has signed with restaurant-technology provider OneDine LLC to deploy its platform in …

Digital Transactions