Okta has made Okta for AI Agents generally available, offering identity, access management and governance capabilities for enterprise AI agents operating across frameworks, cloud platforms and SaaS applications.
Okta has made Okta for AI Agents generally available, introducing an identity-management platform designed to help organisations discover, onboard, protect and govern AI agents across enterprise environments.
In its announcement, Okta said the platform gives AI agents “first-class identities.” The company said those identities are intended to work across agent frameworks, cloud platforms and software-as-a-service applications, allowing organisations to apply identity and access controls to software that can perform tasks on their behalf.
The release comes as businesses experiment with agents that can retrieve information, call application programming interfaces and execute actions in internal systems. Those capabilities create an additional governance challenge because agents may operate with delegated permissions or access credentials while performing work for employees, teams or business applications.
Okta said its platform is designed to support the discovery and onboarding of AI agents, as well as their ongoing protection and governance. The company’s framing treats agents as entities that should be included in an organisation’s identity-management processes rather than handled as isolated automations.
That distinction is relevant to organisations that may not have a complete inventory of the agents operating in their environments. Agents can be created through formal development projects, embedded in business software or configured by individual teams. Without a way to identify them and assign ownership, security teams may struggle to determine what an agent can access or which permissions should be withdrawn when it is no longer needed.
TechRadar reported that Okta’s framework is intended to discover, register and manage AI agents, including unauthorised or poorly documented “shadow” agents. The publication also reported that the platform includes standardised access controls and a central mechanism for disabling agents.
The product reflects a broader change in enterprise identity and access management. Conventional systems have generally focused on employees, customers, contractors, applications and service accounts. AI agents add another category of software: systems that may make decisions, use delegated permissions and initiate actions across multiple services.
Okta’s announcement presents agent identity as both an authentication and governance issue. Authentication can establish which agent is making a request, while access policies can limit the systems and data that agent is allowed to use. Lifecycle controls can also help organisations register an agent, assign responsibility for it and remove its access when circumstances change.
The identity challenge posed by AI agents has also been discussed by Techzine, whose analysis described the technology as a new identity puzzle for enterprise security teams. The central issue is accountability: organisations need to know which software is acting, on whose behalf it is operating and what boundaries apply to its actions.
Okta for AI Agents is generally available, according to Okta. The company is targeting organisations that want to deploy AI agents while extending existing practices around identity, least-privilege access and governance.
The practical value of the platform will depend on how comprehensively organisations can discover agents, connect them to existing identity systems and enforce policies across different frameworks and applications. It will also depend on whether teams can keep agent permissions aligned with changing business requirements.
The release nevertheless signals that identity providers increasingly see AI agents as entities requiring explicit access management. As agents gain the ability to act across enterprise systems, establishing who they are, what they can do and how their access can be revoked is becoming a distinct security responsibility.
The release comes as businesses experiment with agents that can retrieve information, call application programming interfaces and execute actions in internal systems.
Those capabilities create an additional governance challenge because agents may operate with delegated permissions or access credentials while performing work for employees, teams or business applications.
Discovery and lifecycle management Okta said its platform is designed to support the discovery and onboarding of AI agents, as well as their ongoing protection and governance.
Continue reading
