TLDR: Google AI has launched the Agent Payments Protocol (AP2), an open standard designed to enable secure and interoperable AI agent-led payments across various merchants and digital wallets. This initiative, supported by over 60 organizations including major payment and tech companies like Coinbase, Ant International, American Express, and Mastercard, aims to address critical concerns around authorization, authenticity, and accountability in AI-driven commerce. AP2 utilizes ‘Mandates’ – cryptographically signed digital contracts – to ensure verifiable proof of user instructions, fostering trust and preventing fragmentation in the evolving AI payments ecosystem.
Google AI has officially introduced the Agent Payments Protocol (AP2), an open standard poised to revolutionize how artificial intelligence agents conduct transactions across diverse merchant platforms and digital wallets. Announced on September 16, 2025, this protocol is a collaborative effort with over 60 leading organizations in the payments and technology sectors, including prominent names such as Adyen, American Express, Ant International, Coinbase, Etsy, Forter, Intuit, JCB, Mastercard, Mysten Labs, PayPal, Revolut, Salesforce, ServiceNow, UnionPay International, and Worldpay.
The core objective of AP2 is to establish a common, secure foundation for AI agents to initiate and complete payments on behalf of users. As AI agents become increasingly capable of autonomous actions, the traditional assumption of a human directly clicking ‘buy’ on a trusted surface is challenged. AP2 directly addresses the critical questions arising from this shift:
1. Authorization: Ensuring verifiable proof that a user has granted an agent specific authority to make a particular purchase.
2. Authenticity: Allowing merchants to confirm that an agent’s request accurately reflects the user’s true intent.
3. Accountability: Defining responsibility in instances of fraudulent or incorrect transactions.
According to a Google company blog post, ‘AI agents are capable of transacting on behalf of users, which creates a need to establish a common foundation to securely authenticate, validate and convey an agent’s authority to transact.’ The protocol aims to prevent a fragmented ecosystem by providing a shared language for secure and compliant transactions between agents and merchants.
AP2 functions as an extension of existing frameworks like the Agent2Agent (A2A) protocol and the Model Context Protocol (MCP). It is designed to be payment-agnostic, supporting a wide array of payment types, from traditional credit and debit cards to stablecoins and real-time bank transfers. This flexibility ensures a consistent, secure, and scalable experience for all stakeholders, while also providing financial institutions with the necessary clarity for risk management.
A key innovation within AP2 is the use of ‘Mandates.’ These are tamper-proof, cryptographically-signed digital contracts that serve as verifiable proof of a user’s instructions. Signed by verifiable credentials (VCs), Mandates act as foundational evidence for every transaction. They cater to two primary scenarios: real-time purchases where a human is present to approve specific items, and delegated tasks where users pre-authorize agents to make purchases under defined conditions. This sequence from intent to execution creates an auditable trail, enhancing transparency and reducing errors.
Also Read:
- Ohio State University Launches Programs to Equip Small Businesses with AI Capabilities
- YouTube Unleashes Advanced AI Creation Tools for Shorts Creators
James Tromans, Head of Web3 at Google Cloud, highlighted the protocol’s comprehensive design, stating to Fortune, ‘We built it from the ground up with both existing payment system capabilities in mind, as well as future ones like stablecoins.’ Google has expressed its commitment to evolving AP2 through an open, collaborative process, including engagement with standards bodies, and invites the broader payments and technology community to contribute to this future of AI-driven commerce.
