GoDaddy, Infoblox, Google, and Cloudflare Innovate with Secure Transactions for AI Agents

Innovative Security for AI Agent Transactions

In a significant demonstration of technological advancement, GoDaddy, Infoblox, Google, and Cloudflare came together to showcase how AI agents can securely execute transactions across various online platforms. This milestone was presented at GovWare 2026 during Singapore's International Cyber Week.

The central piece of this initiative is the Chain of Trust, an open reference architecture that facilitates the identification and authentication of AI agents throughout a transaction process. This architecture is a result of collaboration with Singapore's Infocomm Media Development Authority (IMDA) and is based on their Model AI Governance Framework for Agentic AI (v1.5).

As the utilization of AI agents expands in commerce, where they are already assisting individuals and businesses in product searches and purchases, the need for secure and trustworthy transaction processes has become critical. According to a report by McKinsey & Company, the potential for AI agents is staggering, with the capacity to facilitate between $3 trillion to $5 trillion in global consumer purchases by the year 2030.

However, the prevalent challenge lies in establishing trust across different platforms. Currently, many platforms verify agents independently through closed systems, leading to a lack of transparency about the verification process. This ultimately hampers the creation of a reliable trust framework across the AI ecosystem.

The Chain of Trust Explained

The Chain of Trust operates on the principles of the open web, integrating layers of controls built upon established internet standards. This allows organizations to verify an agent at every phase, from discovery to payment and audit across platforms. Here’s how it works:

1. Discovery: Through protocols like Agent Resources Discovery (ARD) or DNS-AID, organizations can locate which AI agents are available and where.
2. Identity: AI agents have a registered identity that anyone can cross-verify with public evidence rather than relying solely on session-based certificates.
3. Authorization: This establishes who manages an agent and its permissions, utilizing OAuth standards.
4. Communication: AI agents can interact with tools and one another via Agent2Agent (A2A) and the Model Context Protocol (MCP).
5. Payment: The process includes the person’s signed authorization and the agent’s identity during checkout through a system called AP2.
6. Traceability: It ensures that all transactions are recorded in append-only logs, complete with a signed receipt for offline verification.
7. Trust: Each agent is assessed based on verified public attestations, providing a score that reflects trustworthiness.

Demonstration at GovWare

The showcasing of this innovative framework included a simulated purchase of a flight from Madrid to Singapore, executed among AI agents that were unfamiliar with each other’s systems. The example demonstrated various components working in tandem. For instance, a mock buyer agent operated within Amazon Web Services, while the seller agent functioned via Cloudflare Workers. The approval and verification process for the transactions were managed through Google’s platform.

A crucial aspect involved a person approving a simulated purchase via a password, while an independent service issued signed mandates via AP2, detailing what the mock buyer agent could purchase and from whom. The architecture was adept at ensuring security, even preventing attempts from rogue scenarios such as look-alike sellers and fake credentials from proceeding.

Industry Collaboration and Future Vision

The execution of this demonstrative transaction highlighted the importance of collaboration across sectors. IMDA’s Lee Wan Sie emphasized the necessity of a trusted and interoperable foundational system for agentic AI deployment on a large scale. He reiterated that robust identity frameworks are integral to achieving this.

According to Jared Sine, Chief Strategy and Legal Officer at GoDaddy, the aim is not to overhaul the internet for AI agents but to extend existing open standards that have been connecting the internet effectively for years. By establishing a verifiable identity for agents, it allows transactions to be executed safely across platforms.

The industry leaders encourage DNS operators, identity providers, payment networks, and cloud platforms to adopt and expand this architecture. They have provided a technical case study titled "Implementing the Chain of Trust for an Agentic Transaction" to detail this innovation. Organizations interested in leveraging this architecture can find more information at openagentcommerce.org.

Conclusion

The collaboration between GoDaddy, Infoblox, Google, Cloudflare, and IMDA marks a significant leap towards creating a secure, transparent, and trustworthy environment for AI-driven transactions across the internet. As AI agents increasingly permeate commerce, such foundational frameworks will be pivotal in promoting innovation and ensuring user confidence in automated systems.

For further details about each company involved in this initiative: