Okta executives detailed a new strategy to secure artificial intelligence agents during a July 14, 2026, technical session. The company will extend its identity platform to treat AI agents as first-class identities. This framework implements controls for discovery, registration, entitlements, and access reviews.
The plan addresses rising enterprise risks associated with agentic AI through comprehensive lifecycle management. Key defenses include least-privilege access and a kill switch for suspicious behavior. Okta also introduced dynamic just-in-time permissions to manage agents throughout the authorization process.
Securing AI agents represents the first of five pillars in Okta’s 2026 product strategy. The company’s Identity Governance platform has surpassed 2,000 customers.