The race to define how artificial intelligence agents will safely handle online transactions is accelerating, with major tech and payments firms stepping in to shape the rules. Google and Mastercard are the latest heavyweights to support efforts led by the FIDO Alliance to establish trusted frameworks for what is increasingly being called “agentic commerce.”
As AI-powered assistants evolve from simple tools into autonomous actors capable of making purchases and managing tasks, the need for reliable infrastructure has become urgent. Several companies—including Visa and Stripe—have already been working on open systems designed to allow these agents to execute payments on behalf of both consumers and businesses.
A New Phase for Digital Authentication
The FIDO Alliance, originally founded in 2012 to eliminate passwords in favor of device-based authentication, is now shifting its focus toward a more complex challenge: ensuring AI agents can act securely and transparently on behalf of users. Its latest initiative involves the creation of an Agentic Authentication Technical Working Group aimed at setting standards for how AI-driven transactions should be verified and authorized.
Central to this effort is a newly contributed framework from Google, known as the Agent Payments Protocol (AP2). By donating this protocol to the alliance, Google is positioning it as a shared, open foundation that can be adopted across the industry rather than controlled by a single company.
Building Trust in AI-Driven Transactions
Alongside Google’s contribution, Mastercard is introducing its own system called Verifiable Intent. Together, these technologies are designed to ensure that AI agents operate within clearly defined boundaries, accurately reflecting user instructions while maintaining strong security safeguards.
The FIDO Alliance has identified three primary focus areas for its work: confirming that user instructions are authentic and verifiable, establishing secure authentication methods for AI agents, and enabling trusted delegation so agents can complete transactions responsibly.
According to Andrew Shikiar, CEO of the alliance, AI agents are rapidly becoming embedded in everyday digital experiences—from shopping online to handling routine administrative tasks. However, widespread adoption will depend on whether users can trust these systems to act in their best interest without errors or misuse.
Industry Collaboration Expands
To drive this initiative forward, the working group includes participation from a wide range of major organizations. Leadership roles are being filled by representatives from Google, OpenAI, and CVS Health, while companies such as Amazon and Okta are also contributing in key supporting roles.
At the same time, the alliance’s Payments Technical Working Group—co-chaired by Mastercard and Visa—is developing complementary specifications focused specifically on financial transactions initiated by AI agents.
Toward an Open Standard for Agentic Commerce
Google has emphasized that transferring AP2 to the FIDO Alliance ensures the protocol remains open, platform-neutral, and guided by community input. This approach is expected to accelerate adoption by giving companies a shared blueprint for integrating AI-driven payments into their platforms.
With contributions from both Google and Mastercard forming the backbone of these efforts, the initiative signals a broader industry commitment to making agentic commerce secure, interoperable, and widely accessible.
As AI agents move closer to mainstream use, the success of these standards could determine how confidently users and businesses embrace the next generation of digital commerce.
