Okta extends security for AI agents with support for all identity providers

Okta is announcing a major push into security and identity management for AI agents. The company is expanding its Okta for AI Agents platform with support for more agent ecosystems, integration with Amazon Bedrock AgentCore, and compatibility with identity providers outside of the Okta ecosystem.

The goal is to help companies gain better control over rapidly growing AI environments where autonomous AI agents are gaining increasing access to sensitive systems, APIs, and corporate data.

AI agents create new security challenges

As agent-based AI gains greater traction, so does the need for identity management and security controls. AI agents can now automate workflows, make decisions, and interact with multiple systems simultaneously, often with limited human oversight.

According to Octa Many organizations today lack sufficient visibility into which AI agents are being used, what resources they have access to, and what actions they can perform.

“Security and IT leaders need a better way to understand where their AI agents are, what they can connect to, and what they can actually do,” said Ely Kahn, chief product officer at Okta.

At the same time, several industry reports indicate that AI agents are rapidly becoming a new attack surface for cyberthreats. Identities linked to machines and AI systems are growing significantly faster than traditional user accounts, creating new risks around over-authorization and lack of control.

Integration with Amazon Bedrock AgentCore

One of the biggest news is the integration with Amazon Bedrock AgentCore, AWS’s platform for generative AI and AI agents. The integration enables managing the identity lifecycle of AI agents built and deployed on AWS.

The features include:

• Discovery of AI agents in enterprise environments
• Registration and identity management of AI agents
• Centralized control of access policies
• Control over which resources agents are allowed to use
• Ability to quickly shut down compromised or unauthorized AI agents
• Logging and telemetry for incident management and compliance

Okta states that the solution is platform-neutral and will also support ecosystems such as Salesforce Agentforce, ServiceNow AI Platform, Google Vertex AI, Boomi, DataRobot, and Workday.

Support for all identity providers

Another important part of the initiative is to Okta for AI Agents can now be used with other identity providers, such as Microsoft Entra ID and Ping Identity, meaning that companies do not need to replace their existing identity infrastructure to use the platform.

Instead, organizations can maintain their current human identity solutions while using Okta to secure and govern AI agent identities and access.

It provides companies with a common control plan to:

• Discovering AI agents
• Manage access to SaaS apps and APIs
• Control permissions
• Monitor activities
• Prevent unauthorized access

AI identity will be the next big area of security

The development shows how identity security is quickly becoming one of the most important areas in AI and cybersecurity. Several security experts are now warning that AI agents risk becoming ”new unmanaged endpoints” if companies lack clear governance and policies.

At the same time, the need for vendor-independent solutions is growing as companies use multiple AI platforms in parallel.

Octa is now clearly positioning itself in this space by offering a neutral identity platform for AI agents that works across multiple clouds, ecosystems and identity providers.

About Okta

Octa is an American cybersecurity and identity company specializing in identity and access management for enterprises and developers. The company is increasingly focusing on AI security, agentic AI, and non-human identity management.