Three major developments are converging to solve a fundamental problem in agentic commerce: how can an AI agent act as your proxy to make purchases, access accounts, and complete transactions without becoming a security liability? The answer involves a carefully orchestrated separation of concerns, where different companies handle authentication, authorization, and payment in ways that keep your secrets safe while letting machines do the work.

Why AI Agents Need Their Own Identity Layer?

When Claude, ChatGPT, or any other AI agent reaches a login page, a critical question emerges: what identity is it acting under, and what access should it get? The traditional answer has been unsatisfying. Either you hand over your password to the AI, which is a security nightmare, or you stop and do the task yourself, which defeats the purpose of having an agent in the first place.

"Neither is the future we should build toward," according to 1Password's approach to solving this problem. The company has built a zero-exposure architecture for Anthropic's Claude that allows the AI to complete browser tasks requiring logins and one-time passcodes, but the credentials never enter the model or its memory.

"It is to let a user give an agent permission to use a credential without letting the agent see it. Claude knows it used your login; it does not need the password or one-time code in its context. That distinction is where trust in agents starts and the foundation we're building with Anthropic," said Nancy Wang, CTO of 1Password.

Nancy Wang, CTO at 1Password

In practice, here's how it works: when Claude needs to sign in, 1Password shows the user which credential is being requested and why. After the user gives biometric approval, 1Password injects the credential directly into the page. Claude never sees the vault item, password, or one-time code. Access is scoped to the current task and ends when the task is complete.

How Are Companies Building the Delegated Authority Stack?

The emerging solution involves three separate layers, each addressing a different part of the problem:

• Authentication Layer: 1Password's zero-exposure architecture ensures agents can access accounts without ever seeing credentials, using biometric approval and runtime-only credential injection.
• Authorization Layer: Dai Nippon Printing's CATRINA platform uses verifiable credentials and distributed identity management to confirm that an AI agent is acting as the user's legitimate proxy with appropriate authority.
• Payment Layer: The Linux Foundation's newly operational x402 Foundation stewards an open standard that embeds secure payment capabilities directly into web interactions, allowing AI agents and APIs to send and receive payments as seamlessly as they exchange data.

Dai Nippon Printing has begun providing digital identity management for AI agents making proxy purchases in online shopping. The system leverages verifiable credentials to "create an environment where only AI agents authorized by the user can conduct transactions based on the user's will and authority," according to the company's announcement.

The firm emphasizes that agentic commerce has created a new challenge: confirming and verifying the user to determine whose proxy the agent is acting as and with what authority. By extending identity verification and access control technologies to AI agents, DNP provides a foundation for safe, transparent, and fair online transactions.

What Is x402 and Why Does It Matter for AI Commerce?

The x402 protocol represents a fundamental shift in how payments work on the internet. Rather than treating payments as a separate, manual process, x402 embeds secure payment capabilities directly into web interactions so that AI agents, APIs, and applications can send and receive payments as seamlessly as they exchange data.

The protocol supports payment types ranging from traditional credit cards to stablecoins, making it flexible enough to handle both conventional finance and emerging digital currencies. The Linux Foundation now governs the x402 Foundation, an open-governance body created by Coinbase to steward the protocol's development.

"AI agents and automated systems are becoming active participants in the global economy, yet they have lacked a native, secure way to transact. The operational launch of the x402 Foundation marks a vital milestone in establishing an open, community-governed standard for payments over HTTP," said Jim Zemlin, CEO of the Linux Foundation.

Jim Zemlin, CEO at the Linux Foundation

The Foundation already boasts 40 member organizations, including major players like Adyen, Amazon Web Services, American Express, Circle, Cloudflare, Coinbase, Fiserv, Google, Mastercard, Ripple, Shopify, Solana Foundation, Stellar Development Foundation, Stripe, and Visa. This breadth of support signals that the industry is converging on a common standard rather than fragmenting into competing proprietary systems.

How Do These Three Pieces Work Together?

The three developments illustrate how emergent authentication efforts can harness the power of AI without sacrificing customer agency, privacy, or security. If DNP addresses an agent's authority to buy and 1Password addresses its authority to access the account needed to buy, x402 addresses the transaction itself.

Together, they separate the person, the agent, the authority delegated to it, the credentials it may use, and the transactions it can execute into distinct, manageable pieces. This separation is crucial because it means no single point of failure can compromise everything at once.

When you authorize an AI agent to book a flight, for example, the agent doesn't need to know your password. It doesn't need to know your credit card number. It doesn't even need to know the full scope of your financial authority. Instead, it receives a narrowly scoped delegation that says: "You can book one flight on this airline, up to this price, on this date." The agent executes that task, and the delegation expires when the task is complete.

This architecture represents a fundamental rethinking of how trust works in digital systems. Rather than trusting the agent with your secrets, you're trusting a carefully designed system where each component has only the information it needs and the authority it requires. The agent becomes a tool you control, not a proxy you have to blindly trust with your most sensitive information.