Okta unveils AI Agent governance tools for secure runtime access

Okta announced new capabilities for its Okta for AI Agents platform aimed at controlling and securing the runtime behavior of artificial‑intelligence agents in enterprise environments. The features include an Agent Gateway, Agent‑to‑Agent Connections and Resource Access Certifications that allow companies to apply real‑time policies, enforce least‑privilege access, and continuously validate agents as they interact with corporate applications and each other.

At the Okta AI Identity Summit in Frankfurt, Jan Brose highlighted how the solution meets German compliance requirements such as NIS2, GDPR and DORA, while Chief Product Officer Ely Kahn stressed the importance of runtime governance for complex, high‑privilege AI workflows. The tools are designed to register AI agents as identities, manage their lifecycle, and disable unauthorized agents, supporting secure, agent‑driven business processes.