Mastercard & Google Pilot AI Agents for Payments
Mastercard and Google are partnering on "verifiable intent" technology as the first AI agents begin making payments on behalf of users. The move aims to build a new standard for security and trust in autonomous transactions, a critical step for the future of AI-driven commerce and fintech.
The core of the new system is a tamper-resistant, cryptographic record that links the user's identity, their specific instructions to the AI, and the resulting transaction details. This creates a verifiable audit trail that can be used by the consumer, merchant, and card issuer to resolve any disputes, addressing the critical question of proving an AI agent acted as instructed. This "Verifiable Intent" framework is built upon existing open standards from bodies like the FIDO Alliance, EMVCo, and the World Wide Web Consortium. While co-developed with Google and aligned with its Agent Payments Protocol (AP2), the technology is designed to be protocol-agnostic, allowing it to function across different AI platforms, digital wallets, and even other payment networks. Mastercard's Chief Digital Officer, Pablo Fourez, emphasized that as AI autonomy in commerce increases, user trust can no longer be implied but must be proven with verifiable evidence. The system uses selective disclosure, meaning only the minimum necessary transaction data is shared for fraud checks or dispute resolution, preserving user privacy. This initiative is part of Mastercard's broader "Agent Pay" program, which has already seen pilot transactions. In early March 2026, Banco Santander and Mastercard completed Europe's first live, end-to-end payment by an AI agent using this infrastructure. Previous trials have also been conducted with partners like Citi and U.S. Bank. The collaboration aims to establish an industry-wide trust layer before agentic commerce scales massively. Beyond Google, Mastercard is working with other key players like IBM and Worldpay to drive adoption. The specifications and a reference implementation have been open-sourced on GitHub to encourage broad integration by developers and partners. This move addresses a fundamental shift from simple e-commerce chatbots to autonomous AI agents that can independently make purchasing decisions based on user-defined goals, like reordering groceries or booking travel within a budget. The goal is to create a secure and trusted environment for this next generation of AI-driven commerce to scale globally.
