TLDR: Major payment industry players, including Google, Visa, and Mastercard, have unveiled new protocols and frameworks designed to establish secure and trustworthy transactions for the rapidly evolving AI agent economy. Google introduced its Agent Payments Protocol (AP2), utilizing cryptographically signed “Mandates” for auditable transactions. Visa launched its Trusted Agent Protocol, focusing on verifying AI agents for merchants and integrating with stablecoins. Mastercard debuted its Agent Pay program, featuring Agentic Tokens and collaborations with tech giants to ensure secure, transparent, and consumer-controlled AI-driven payments.
The advent of artificial intelligence agents capable of autonomous actions, including monetary transactions, is ushering in a new era of commerce. This shift necessitates a robust and secure payment infrastructure that can operate without the traditional “human click.” In response, leading financial and technology companies have introduced groundbreaking protocols to ensure trust, authenticity, and accountability in the burgeoning agentic economy.
Google’s Agent Payments Protocol (AP2): A Global Rulebook for AI Commerce
Google, in collaboration with over 60 major payment providers and tech companies, has developed and announced the Agent Payments Protocol (AP2). This open protocol aims to establish a common language for safe, interoperable transactions in AI-driven commerce. AP2 addresses three critical challenges: authorization (proving an agent is permitted to make a specific transaction), authenticity (guaranteeing the agent’s request reflects the user’s true intent), and accountability (clarifying responsibility in case of fraudulent or erroneous transactions).
At the core of AP2 are “Mandates,” cryptographically signed, tamper-resistant digital contracts that serve as verifiable proof of a user’s instructions. These include “Intent Mandates” for initial requests and “Cart Mandates” for user-approved purchases, creating a non-repudiable audit trail. The protocol also emphasizes role separation, with entities like Shopping Agents, Merchant Endpoints, Credential Providers, and Payment Processors exchanging Verifiable Credentials (VCs) to ensure security without the Shopping Agent directly accessing payment credentials. AP2 is designed to support diverse payment methods, including stablecoins and cryptocurrencies, and has an A2A x402 extension developed with partners like Coinbase, the Ethereum Foundation, and MetaMask to accelerate Web3 ecosystem support.
Visa’s Trusted Agent Protocol: Verifying Intent in AI-Driven Transactions
US-based digital payments firm Visa has launched its Trusted Agent Protocol, a framework designed to facilitate secure transactions between merchants and AI agents. Developed in partnership with Cloudflare and other key players such as Ant International, Adyen, Checkout.com, CyberSource, Coinbase, Elavon, Fiserv, Microsoft, Nuvei, Stripe, Shopify, and Worldpay, this protocol addresses the surge in AI-driven traffic to retail websites, which has increased by over 4,700% in the last year. Merchants face challenges in distinguishing legitimate AI agent transactions from harmful bots and maintaining visibility into the actual consumer.
The Trusted Agent Protocol allows authorized agents to transmit essential information to merchants, leveraging agent-specific cryptographic signatures. This information includes the agent’s intent, consumer recognition details, and payment information, supporting various checkout methods. Jack Forestell, Visa’s chief product and strategy officer, stated, “We believe the entire payments ecosystem has a responsibility to ensure sellers can trust AI agents as much as they trust their best customers and networks.” He added that the protocol focuses on “creating no-code functionality for merchants to securely identify agents with an intent to buy and provide a better payments and personalised experience for its known users.” The protocol is based on the HTTP Message Signature standard and aligns with Web Both Auth, also working to align with other protocols like the Agentic Commerce Protocol.
Mastercard’s Agent Pay: Secure and Personalized AI Commerce
Mastercard has unveiled its Agent Pay program, an agentic payments technology designed to revolutionize commerce in the age of AI. This solution aims to deliver smarter, more secure, and more personal payment experiences for consumers, merchants, and issuers. The program introduces “Mastercard Agentic Tokens,” which build upon existing tokenization capabilities to enable secure payments initiated through conversational interfaces.
Also Read:
- Paid Secures €10 Million to Propel AI Agent Economy and Beta Program Launch
- Pinterest Unveils AI Shopping Assistant and Enhanced Generative AI Controls for Personalized Discovery
Mastercard is collaborating with Microsoft on new use cases, integrating technologies like Microsoft Azure OpenAI Service and Microsoft Copilot Studio to scale agentic commerce. Partnerships with technology enablers like IBM (with its watsonx Orchestrate product) and checkout players like Braintree and Checkout.com are also in place to enhance tokenization for transparent agentic payments. The program emphasizes securely registering and authenticating trusted AI agents, establishing clear rules for consumer control over purchases, and protecting against fraud through Mastercard’s robust cybersecurity and authentication capabilities. Jorn Lambert, chief product officer at Mastercard, highlighted the company’s commitment to “redefining commerce in the AI era” and advancing standards for agentic payments, such as applying the Model Context Protocol to Secure Remote Commerce.
