Okta announced a set of innovations to secure the AI-driven enterprise and combat related threats, Okta said on 25 September 2025. The release ties identity controls more tightly to AI agent and application access patterns.
As companies connect models and agents to corporate data, identity becomes the enforcement point for who — or what — can call which API. Vendors that already own SSO and lifecycle management are extending those graphs into machine and agent identities.
Okta’s Auth0 heritage and workforce IAM footprint give it distribution into the same two buyer groups now writing AI access policies. The September pack is part of that repositioning.
Security architects should ask which controls are generally available versus preview, and how agent identities are governed alongside human users, before counting the announcement as a completed control upgrade.