Agentic Identity.
The identity model for AI agents acting autonomously or on behalf of users, a third category of identity alongside human users and traditional non-human (service) identity, with its own authentication, authorization, and audit patterns.
Agentic identity is the third leg of the identity stool: human (CIAM), workload (NHI), agent. Conflating any two of these is the root of the worst agentic-AI security incidents: an agent running with a user's full token (treating the agent as the user) means the agent's bugs become the user's compromises; an agent running as a generic service account (treating the agent as a workload) loses the audit trail of which user requested which action.
That split is now blurring at the category level. Gartner's 2026 Magic Quadrant for CIAM defines the category as access for "customers, partners and machine identities," adding that "these identities include AI and can exist across all constituency groups" (definition as quoted in Okta's announcement, citing the report dated October 28, 2026). Ping Identity's announcement describes the report as a dedicated CIAM evaluation, distinct from the broader Access Management scope assessed in 2025.
What machine identity inside CIAM means (our reading). It is not workload identity for your own back end, which stays with NHI tooling. It is the agent or machine client that acts for a customer or partner: an assistant buying on a customer's behalf, or a partner's integration calling your API. The CIAM system has to register that client, bind it to the customer or partner who delegated to it, scope what it may do, and log both parties on every action. That is the same delegation record Know Your Agent depends on.
Common questions
What's the difference between agentic identity and NHI?
How do AI agents authenticate to APIs?
Why does agentic identity need its own discipline?
Related terms
In the guides
Authentication for AI Agents: OAuth Patterns for Non-Human Identity
How AI agents authenticate in 2026. The on-behalf-of pattern, delegated agent identity, OAuth 2.1 Dynamic Client Registration, and where the patterns are still being invented.
Authorization Patterns for Agentic Workflows: Delegation, Constraints, and Just-in-Time Permissions
AI agents need authorization models that handle delegated permissions, multi-step workflows, and least-privilege at machine speed. The patterns that work and the ones being invented.
MCP Server Identity Model: Authentication, Authorization, and Trust for the Model Context Protocol
MCP is OAuth 2.1 with discovery. How MCP clients identify themselves (CIMD first, DCR as fallback), how servers scope access, and what the spec leaves to you.
Token Management for AI Agents: Lifetimes, Rotation, and Revocation at Machine Speed
Agent tokens are stolen faster and used harder than human tokens. How to set lifetimes, rotate refresh tokens, scope per-tool, and detect anomalies in production agent deployments.