AI agents are starting to show up everywhere. They answer questions, interact with software, and increasingly perform tasks on behalf of users. As those agents begin talking to other systems and even other agents, a new problem is becoming obvious: how do you know an AI agent actually belongs to the company it claims to represent?
That is the issue behind a new initiative from GoDaddy. The company announced a public implementation of the Agent Name Service, or ANS, which is designed to give AI agents verifiable identities tied to the infrastructure that already powers much of the internet.
Rather than inventing an entirely new system, ANS builds on familiar technology. Each AI agent is associated with a human readable name tied to a domain, and that identity can be verified using DNS records along with cryptographic certificates. In theory, this allows software systems to confirm where an AI agent came from before interacting with it.
The approach is somewhat similar to how websites prove their authenticity using encrypted connections and certificates. Instead of verifying a website, however, the system would verify the origin of an AI agent.
GoDaddy says its implementation, the unimaginably named “GoDaddy ANS,” allows organizations to publish and verify AI agents in minutes. Once registered, an agent becomes discoverable through DNS and can be verified by both humans and other systems.
The first company to test the system publicly is LegalZoom. The company has registered an AI agent that integrates its legal services into AI assistants. That agent can help users connect with attorneys, share documents, and manage consultations through conversational interfaces.
In practice, the idea is that an AI assistant could discover the LegalZoom agent and confirm through DNS that it actually belongs to LegalZoom before interacting with it.
Supporters say this kind of identity layer could become necessary as AI agents spread across the web. Without a reliable way to confirm ownership, rogue agents could impersonate legitimate companies or trick users into sharing sensitive information.
At the same time, the concept may raise questions among developers and open web advocates. If verifying AI agents depends on domain ownership and DNS infrastructure, domain registrars and certificate authorities could end up playing a central role in how AI agents are trusted online.
In other words, the same system that helps your browser find a website might also become part of how AI systems determine which agents are legitimate.
Whether ANS becomes widely adopted remains to be seen. Engineers may appreciate the decision to build on existing internet standards instead of introducing yet another identity framework. But as with many proposals tied to the emerging “agentic web,” the long term impact will depend on whether developers and platforms actually choose to support it.
For now, GoDaddy is clearly betting that AI agents will eventually need a trust layer built on the same infrastructure that has kept the web functioning for decades. LegalZoom’s early participation suggests at least some companies are willing to experiment with that idea.
Support independent tech journalism
NERDS.xyz is independently owned and operated. If you enjoy my coverage of Linux, AI, hardware, cybersecurity, and tech culture, consider supporting the site on Ko-fi.
Support NERDS.xyz