Okta Enhances AI Agent Management Amid Rising Non-Human Identities

The enterprise AI sector is changing quickly, with Gartner predicting that by 2028, the average Fortune 500 company will deploy over 150,000 AI agents. This surge in non-human identities requires effective governance and management solutions, a challenge that Okta is tackling with updates to its Okta for AI Agents product.

In a strategic move, Okta has integrated Amazon Bedrock AgentCore into its platform. This enables organizations to manage AI agents across various identity providers, not limited to Okta’s own. The integration offers advanced controls over agent identities built on Amazon Web Services, including ownership assignments and lifecycle management tools. Furthermore, the ability to discover agents through OAuth consent monitoring and import them via the Okta Integration Network enhances the platform's functionality.

As AI deployments become more complex, Okta’s enhancements equip clients with tools to define governance policies centrally. Registered agents can be linked to human owners, allowing organizations to specify resource access, authentication methods, and action scopes. This is particularly important considering that Okta reports 90% of enterprise agents are over-permissioned, with 53% having access to sensitive data. This trend highlights the urgent need for improved visibility and control over AI agent activities.

The updated product features workflow tools for managing user access requests, enabling clients to revoke an agent's access with a single command. Additionally, comprehensive logging of system interactions aids compliance and incident response efforts, ensuring organizations remain accountable as they scale their AI operations.

Okta's platform is well-suited to manage AI agents across mixed technology environments. The recent updates support various agent builders, including Salesforce Agentforce and ServiceNow AI Platform, allowing companies to leverage their existing infrastructure while benefiting from Okta’s governance capabilities.

The second phase of Okta's update introduces support for non-Okta identity providers. This move facilitates the integration of legacy systems like Microsoft Entra ID and Ping for human user identities, allowing organizations to maintain their existing identity frameworks for employees while using Okta to oversee AI agent lifecycles. This effectively creates a unified control plane for managing access to diverse enterprise resources, including software-as-a-service applications and APIs.

As companies tackle the complexities of an expanding agent environment, Okta's latest advancements meet the urgent demand for effective governance. The expected increase in AI agents presents both a challenge and an opportunity for organizations to rethink their identity management strategies to ensure security and compliance. The future is likely to bring further innovations as firms adapt to the evolving dynamics of AI integration.