Infoblox & GoDaddy back open standards for AI agents

Infoblox and GoDaddy are backing open standards for AI agent discovery, identity and verification on the open web.

The work centres on complementary approaches built on the Domain Name System.

Infoblox is advancing DNS for AI Discovery, known as DNS-AID, while GoDaddy is helping develop Agent Name Service, or ANS.

Both are designed as open standards that would let AI agents identify themselves, be discovered by other systems and provide a way to verify that they are genuine. The aim is to place those functions on existing internet infrastructure rather than on a proprietary registry or a system controlled by a small number of vendors.

The move comes as technology groups and regulators pay closer attention to the risks of autonomous software operating across websites, applications and corporate environments. As AI agents are increasingly used to search, transact and interact on behalf of people or businesses, operators will need a reliable way to know which agent they are dealing with and where it is allowed to connect.

DNS-AID is being developed as an Internet Engineering Task Force draft and as open-source software. It sets out how AI agents can publish metadata in DNS records so other systems can find details such as endpoints and supported protocols.

ANS, also being developed through the IETF process, focuses on naming, identity and verification. It is designed to let operators use domain names they already control, alongside public key infrastructure, instead of relying on a new naming system.

Shared approach

The two efforts are intended to work alongside each other rather than compete. In simple terms, ANS is meant to address who an agent is, while DNS-AID helps other systems find what that agent does and how to reach it.

That distinction matters because the growth of AI agents is expected to raise overlapping technical questions around trust, discovery and governance. A system may need to establish both that an agent belongs to a known operator and that the service details published for it are legitimate.

Jared Sine, chief strategy and legal officer at GoDaddy, said using DNS gives the work an internet-scale foundation. "Agents will only reach their full potential on the open web if people and systems can verify who they are interacting with," he said. "Adopters of the Agent Name Service open standard leverages the only infrastructure that exists today that operates at the scale and speed of the global internet - Domain Name Service. We support Infoblox's work on DNS-AID and believe open standards for identity, discovery and verification will be critical as agents become part of everyday digital experiences."

Infoblox makes a similar case for DNS, pointing to its long-standing role as a distributed naming system for the internet. It argues that the same architecture that supports websites and email can also serve as a trust and discovery layer for software agents.

Wei Chen, CLO and EVP of regulatory strategy at Infoblox, linked the proposal to the history of internet infrastructure. "The lesson we learned from the 1970s-1980s is simple: no single entity could or should run the phonebook of the internet for everyone," Chen said. "DNS replaced it, not with another centralised list, but with an open, federated protocol that anyone could participate in. Forty years later, DNS remains the gold standard for digital trust and a scalable foundation where agents, Model Context Protocols, services and endpoints can be discovered and trusted through the same architecture that already powers the global economy."

Standards debate

The announcement also reflects a broader debate over how AI agents should be governed as they move from experimental tools to internet-facing services. Industry groups are beginning to discuss whether identity and discovery should be embedded in open protocols, left to platform operators or handled through vendor-specific marketplaces and registries.

By tying the work to DNS and PKI, Infoblox and GoDaddy are arguing for a federated model. Under that approach, organisations would retain control over their own agent identities and metadata, while trust decisions would be based on auditable cryptographic signals rather than reputation scores managed by a single provider.

That may appeal to companies wary of allowing one platform to become the default directory for AI agents. It could also align with existing governance models for domain ownership, certificate management and internet standards development.

DNS-AID uses existing DNS record types including Service Bindings, DNS service discovery, DNSSEC and DANE. Relying on established standards is intended to avoid the need for entirely new infrastructure while giving developers and operators a way to publish machine-readable information in a form other internet systems already understand.

ANS follows the same broad logic from the identity side. By letting agent operators use domain names they already own, the proposal aims to make AI agents addressable through the same naming layer used for websites and email.

The companies are calling on cloud providers, agent platform vendors, registrars, security firms and standards bodies to take part. "We are calling on cloud providers, agent platform vendors, registrars, security companies and standards organizations to join us in open standards work," Sine said. "We believe AI agents should be discovered and verified through open infrastructure that is fully federated and distributed."