AI agents are beginning to move from novelty to infrastructure. They will increasingly search, transact, negotiate, book, approve, recommend and act across digital services.
That creates a fundamental trust problem: how do we know which agent we are interacting with, who it represents and what it is authorised to do?
On 23 June, the Linux Foundation announced its intent to launch the Agent Name Service (ANS), an open standard for trusted identity, verification and discovery of AI agents. Built on the existing Domain Name System, ANS is designed to anchor agent identity to infrastructure the internet already relies on.
Rather than creating a new proprietary registry, ANS uses DNS as a neutral foundation. This matters. DNS already operates at global scale and provides a familiar model for naming and discovery. By building on it, ANS points toward a more open and interoperable approach to agent identity.
For DINZ members, several parts of the announcement are particularly relevant.
First, ANS supports decentralised identifiers (DIDs) and Legal Entity Identifiers (LEIs). This creates a potential bridge between agent identity, organisational identity and the verifiable credential ecosystem. In practice, that could help users verify not just that an agent exists, but who it represents and whether it has appropriate authority.
Second, ANS is focused on discovery and verification. In an agentic economy, agents will need to find each other and determine whether interaction is safe. That requires more than a name. It requires assurance, authenticity, permissioning and trust signals that can be checked in real time.
Third, the initiative reinforces the importance of neutral infrastructure. If agent identity becomes controlled by a small number of proprietary platforms, the same risks that exist in today’s platform economy will intensify: lock-in, opacity, concentrated control and weak accountability. Open standards help reduce those risks.
The Linux Foundation’s framing is aligned with a broader theme in digital identity: open standards and shared infrastructure beat walled gardens. For Aotearoa, that matters because agent identity will quickly become part of digital trade, public service delivery, enterprise automation and consumer protection.
There are also sovereignty questions. If AI agents are operating on behalf of New Zealand organisations, communities or individuals, who governs their identity? How are they recognised? How are permissions managed? How do Māori data sovereignty principles apply when machine actors interact with data, services and decisions?
ANS does not answer all of these questions. But it is an important signal that trusted identity infrastructure for agents is becoming a global priority.
For DINZ, the opportunity is to engage early. Aotearoa should not simply inherit agent identity infrastructure designed elsewhere. We should help shape it — with interoperability, privacy, accountability and sovereignty built in from the outset.