New feature of Okta for AI Agents that certifies what your agents can connect to and whether it should still exist, with access revocations enforced.

✦ AI blueprint: What can they connect to?

New feature of Okta for AI Agents that delegates day-to-day AI agent management to a scoped admin role instead of granting Super Admin access, reducing risk. Admins get permissions to create, update, and delete AI agents and MCP servers, configure agent imports for app integrations, and monitor agent activity.

✦ AI blueprint: Where are my agents? What can they connect to? What can they do?

New feature that integrates CrowdStrike Falcon AI Detection and Response (AIDR) with Okta for AI Agents by surfacing local agents installed across employee endpoints, which MCP servers are active and what systems they connect to, who installed each agent, on which device, under which identity, and what each agent can reach.

✦ AI blueprint: Where are my agents?

New SKU that unlocks Okta for AI Agents capabilities* for customers in regulated environments (FedRAMP High, FedRAMP Moderate, and HIPAA).

*Note: This plan does not include shadow AI agent discovery or secure agent connections to service accounts and secrets. We are working on getting these features authorized. Check here for updates.

✦ AI blueprint: Where are my agents? What can they connect to? What can they do?

New feature of Okta for AI Agents that closes the gap in multi-agent architectures by providing an identity layer for agent-to-agent handoffs—designed to provide visibility and audibility for your agent interactions.

✦ AI blueprint: What can they connect to?

Use case implementation of Fine-Grained Authorization and Okta for AI Agents that moves beyond static roles and access tokens to continuous evaluation of your agent action based on real-time context.

✦ AI blueprint: What can they do?

Enables customers to define, audit and secure AI agents as distinct, verifiable machine identities with strict runtime permissions.

Support third-party API integrations and safeguard against cross-tenant data leaks by replacing static, hardcoded credentials with restricted, short-lived tokens.

New integration that lets customers import AI agents built on Amazon Bedrock AgentCore and govern them directly in Okta for AI Agents.

✦ AI blueprint: Where are my agents?

Okta for AI Agents works alongside non-Okta identity providers, giving organizations a purpose-built agent identity platform without requiring them to replace their existing human identity infrastructure.

✦ AI blueprint: Where are my agents? What can they connect to? What can they do?

Optimizes relationship-based access control by indexing permission paths, enabling you to use AI agents to verify document-level user authorization in milliseconds.

Foundational offering that provides the identity layer to discover, onboard, protect, and govern AI agents from a single control plane.

✦ AI blueprint: Where are my agents? What can they connect to? What can they do?

Provides a more secure foundation for developers to build customer facing AI agents.