Okta Expands AI Agent Security to Amazon Bedrock and Any Identity Provider

Okta’s Okta for AI Agents platform now supports Amazon Bedrock AgentCore, Microsoft Entra ID, Ping Identity, and other third-party identity providers, extending the platform beyond Okta-native environments to serve as a vendor-neutral governance layer for AI agent identity across enterprise systems.

The most significant new integration is with Amazon Bedrock AgentCore, Amazon Web Services’ managed generative AI service. Through the Okta Integration Network, customers can import agents built on Bedrock directly into Okta, assign human owners, attach baseline access policies, and log all tool calls and authorization decisions to SIEM systems. Seven capabilities are available for Bedrock-built agents: AI Agent Discovery, AI Agent Import, AI Agent Registry with owner assignment, Resource Connections with authentication method enforcement, User Access Requests and Certifications, Agent Deactivation (a single-action kill switch), and System Logs and Telemetry.

The platform can now also layer governance atop non-Okta identity providers. Enterprises running Microsoft Entra ID or Ping Identity’s own AI agent governance platform can add Okta’s governance controls without replacing their existing workforce identity infrastructure. Planned future integrations include Salesforce Agentforce, ServiceNow AI Platform, Glean, Google Vertex AI, and Workday.

“Security and IT leaders need a better way to understand where their agents are, what they can connect to, and what they can do,” said Ely Kahn, Chief Product Officer at Okta. “The agentic enterprise doesn’t fit neatly into one vendor’s ecosystem. Okta for AI Agents is the only platform built for that reality.”

Okta cited Gartner forecasts that Fortune 500 companies will operate more than 150,000 agents by 2028, alongside Obsidian Security research indicating that 90 percent of enterprise agents are currently over-permissioned and 53 percent access sensitive information. Ric Smith, President of Products and Technology at Okta, noted that “AI agents are evolving faster than any software before them, making traditional security models obsolete.”

The May 14 expansion follows a sequence of product releases. Okta for AI Agents reached general availability on April 30, 2026, following a blueprint announcement at Okta Showcase in March 2026. Earlier in February 2026, Okta introduced Identity Security Posture Management agent discovery features targeting unauthorized or unregistered agent deployments. Okta flagged AI agent security as a strategic priority in its Q3 FY2026 earnings in December 2025. The platform’s multi-IDP positioning also reflects broader ecosystem activity, including World’s proof-of-human integration with Okta in April 2026, demonstrating how the platform is already embedded in third-party agentic trust frameworks.

Sources: Okta Newsroom, Okta Investor Relations, SiliconAngle

–

By the ID Tech Editorial Team