TL;DR: AI agents still need authentication, authorization, and audit foundations just like human users, according to WorkOS. Identity governance has to cover agents and workloads together, or monitoring and guardrails stay separate from revocation and access control.
Editorial analysis by NHI Mgmt Group, based on content published by WorkOS: “Noma Security vs WorkOS: Choosing the Right Platform for Agentic Security”.
Key questions
Q: Why do AI agents increase the need for shared authorization logic?
A: AI agents create more runtime access decisions because they initiate actions across systems without being tied to a single application boundary.
Q: How can security teams tell whether agent access is actually under control?
A: Look for evidence that the team can trace every tool call, secret use, and cross-system action back to a named owner and a valid approval path.
Q: What breaks when AI agents are given access without identity governance?
A: What breaks is accountability.
Practitioner guidance
- Define a single identity path for agents Bind each agent to an initiating user or workload identity so authentication, authorization, and audit all resolve through one governed path.
- Use relationship-based authorization for delegated actions Model what an agent may access through resource relationships and entitlements, then enforce the same policy graph across human and agent requests.
- Align revocation with user offboarding Ensure that terminating a user or service context immediately removes any agent permissions derived from that context, including token and session inheritance.
Bottom line: Agentic security platforms still need identity foundations because monitoring cannot replace authentication, authorization, and revocation.
Glossary terms: Agentic Security Platform, Delegated Identity, Relationship-Based Access, Control Drift
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
Agentic security does not replace identity foundations. The article's central claim is that monitoring, guardrails, and model-centric controls cannot stand in for authentication, authorization, and revocation. That is the right frame for enterprise adoption because every agent still needs a governed identity path before it can safely act. Practitioners should treat AI security as an extension of identity governance, not a substitute for it.
A few things that frame the scale:
- 69% of security leaders agree identity management must fundamentally shift to address agentic AI systems, according to the 2026 Infrastructure Identity Survey.
A question worth separating out:
Q: What should teams do when customer agents act on behalf of users?
A: Enforce blended identity so the agent’s capability is always constrained by the user’s permission set and the tenant boundary. That prevents a shared agent instance from reusing the wrong context across customers and keeps delegated actions auditable at the right scope.
👉 Read our full editorial: Agentic security platforms still depend on identity foundations