Discover and Register AI Agents
This learning path shows Okta administrators how to find AI agents already operating in their environment, register them as managed identities in Okta, and apply the access controls and audit trail needed to govern them safely.
Unregistered AI agents authenticate directly with SaaS platforms and internal APIs, leaving administrators with no owner, no credential, and no way to shut a rogue agent down without breaking the systems it touches. This learning path shows Okta administrators how to close that gap using Okta's four-phase AI agent governance framework.
- Detect shadow AI agents operating outside Okta's visibility using Identity Security Posture Management (ISPM) direct connectors and the Secure Access Monitor (SAM) browser plugin.
- Triage discovers agents and registers legitimate ones as managed Workload Principal identities in Universal Directory.
- Register a custom-built AI agent manually, assign it a human owner, and generate a public key credential to replace static, hardcoded secrets.
- Apply the Principle of Least Privilege by configuring machine-to-machine and cross-app access patterns, and use the Okta System Log to trace every agent action back to an owner and an authorization decision.
Feel like you know your stuff after completing the training in this path? Pass the Assessment to earn a skill badge.
Format: On-demand
Duration: 55 minutes