In a significant move today, Okta Inc. has broadened the capabilities of its Okta for AI Agents platform, enabling it to integrate with Amazon Web Services’ Amazon Bedrock. This expansion will reshape how organizations manage artificial intelligence agents across various ecosystems, marking a shift in identity and access management.
Historically, enterprises have struggled to manage AI agents due to their reliance on single-vendor solutions that limit visibility and control. Okta's integration with Bedrock enables organizations to discover, onboard, protect, and govern agents without being restricted to a single vendor’s infrastructure. Ely Kahn, Chief Product Officer at Okta, captured this shift, stating, "The agentic enterprise doesn’t fit neatly into one vendor’s ecosystem. Okta for AI Agents is the only platform built for that reality, securing the full agent lifecycle — discover, onboard, protect and govern, while meeting customers where they are."
New Features and Capabilities
The integration with Amazon Bedrock introduces several key features that enhance agent governance. Teams can assign identities to Bedrock-built agents, designate human owners, and implement access policies at scale. These capabilities include shadow agent discovery, which monitors OAuth consent grants on managed browsers like Google Chrome, and the ability to import Bedrock agents directly into Okta through its Integration Network.
Additionally, Okta's registry treats every agent as a first-class identity, equipped with baseline security policies. Administrators can define the resources these agents can access, determine the authentication methods they use, and manage the scopes they receive. Notably, a kill switch feature allows for the immediate shutdown of any misbehaving agents, while system logs maintain a comprehensive record of every tool call and authorization decision, ensuring thorough oversight.
Extending Governance Beyond Bedrock
The expansion does not stop at Bedrock. Okta for AI Agents also accommodates integrations with other platforms, such as Salesforce’s Agentforce and ServiceNow’s AI Platform. Future updates are expected to include compatibility with DataRobot, Boomi, Glean Technologies, Google Cloud’s Vertex AI, and Workday, among others.
Okta’s latest announcement also opens its AI agent platform to organizations using different identity providers, such as Microsoft Corp.'s Entra ID and Ping Identity. This flexibility allows companies to layer Okta functionalities over their existing infrastructure without requiring significant changes to their current systems.
Addressing Agent Sprawl
This strategic enhancement arrives at a critical moment as the industry grapples with the complexities of managing agent sprawl. Okta positions its platform as a control plane for agent identities that encompasses software-as-a-service applications, application programming interfaces, and various service accounts. This framing highlights an industry-wide concern: if agent management is treated as an afterthought, governance challenges will escalate.
The introduction of automated workflows for user access requests and certifications, previously reserved for human identities, now extends to AI agents. This stable governance model ensures that access to Bedrock-built agents is streamlined and rigorously monitored, requiring periodic access certification.
Implications for the Future
The implications of Okta’s expansion are substantial. By facilitating a more integrated approach to AI agent management, the company is addressing current governance challenges and anticipating future needs in an increasingly complex technological landscape. As organizations deploy agents across various platforms and resource types, Okta’s initiative could prove pivotal in shaping the future of identity management in the age of AI.
Effectively governing AI agents is becoming essential as enterprises explore artificial intelligence's potential. With Okta leading this transition, the identity management landscape is set for a redefinition that prioritizes flexibility, security, and comprehensive oversight.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.
