Non-human identity (NHI)
A non-human identity (NHI) is a digital identity used by software rather than a person. Learn what counts as an NHI, how AI agent identities fit in, and why these identities need the same governance as human accounts.
What is a non-human identity?
A non-human identity (NHI) is any identity that software uses to authenticate and access resources without a person signing in. Common examples include service accounts, application accounts, API keys and tokens, certificates, cloud workload identities, service principals, scripts, bots, and AI agents.
In most environments, non-human identities outnumber human users. They are often created quickly for an integration or automation, given broad permissions, and then forgotten.
AI agent identity
An AI agent identity is a non-human identity used by an AI agent or assistant that acts on someone's behalf: calling APIs, running commands, or reading data. Agents raise the stakes because they decide which tools to call at runtime. Good practice is to give each agent its own identity, scope it to the task, keep secret values out of the model where possible, and log what it does.
For one example of keeping credentials away from an AI agent, see our documentation: Work with AI agents.
How do non-human identities work?
- An identity is created for a workload, integration, or agent, such as a service account or app registration.
- It receives a credential: a password, key, token, or certificate.
- The software presents that credential to authenticate to a target system or API.
- Permissions attached to the identity decide what the software can do.
- The credential is rotated, and the identity is reviewed or retired when no longer needed.
Why do non-human identities matter?
Non-human identities often hold standing, privileged access, and their credentials can end up hard-coded in scripts or configuration files. Because nobody signs in interactively, misuse can go unnoticed. Treating NHIs as first-class identities, with an owner, a vault for their secrets, least-privilege permissions, rotation, and audit logs, closes a gap that human-focused controls miss.
Frequently asked questions
What is a non-human identity?
A non-human identity (NHI) is a digital identity used by software rather than a person, such as a service account, API key, certificate, workload identity, bot, or AI agent. It authenticates and accesses resources without a human signing in.
Is an AI agent a non-human identity?
Yes. An AI agent that calls tools or APIs needs an identity and credentials, which makes it a non-human identity. Because agents choose actions at runtime, they should get their own scoped identity, limited permissions, and an audit trail.
How do you secure non-human identities?
Inventory them, assign an owner, store their secrets in a vault instead of code, grant least-privilege permissions, rotate credentials regularly, and log their activity so unusual use can be reviewed.
Govern the credentials your workloads use
Store, share, and audit secrets for accounts, scripts, and integrations from a central vault.
Related terms
Service principal
An identity created for an application, service, or automation tool to access resources in Microsoft Entra ID and Azure.
Read now →Secrets management
The practice of securely storing, distributing, rotating, and auditing machine credentials such as API keys, tokens, certificates, and database passwords.
Read now →Identity and access management (IAM)
A system for authenticating and empowering users or non-human identities to seamlessly access the resources necessary to carry out their responsibilities according to the principle of least privilege.
Read now →