Controlling AI Agents with Identity & Access Management

Name: Controlling AI Agents with Identity & Access Management
Uploaded: 2026-06-12T20:32:44-04:00
Duration: 3 min 27 s
Description: TL;DR Okta for AI agents treats autonomous AI systems as first-class identities within the existing Okta platform, bringing enterprise-grade access controls and governance to AI agent deployments. While 91% of organizations use AI agents, only 10% have...

Okta

06/12/2026

0 (0%)

Report Like Favorite

Transcript

That's one of the defining questions in the age of AI agents, especially when you consider this. When you think about an agent, you really think you're really dealing with an entity that has a little bit of reasoning and inference and its own little brain. It's almost like the only human control that our customers have over a process that might just run off and do its own thing is the identity access authorization controls that they can enforce at scale across their enterprise. In this three-part series, we're exploring what it takes to build a more secure, agentic enterprise. In part one, we looked at the growing adoption of AI agents inside the enterprise and security and governance challenges that come with it. Now in part two, we're turning to the solution. In Okta for AI agents, you can really think about it as an extension of our existing Okta platform, but ensuring that AI agents are treated as a first-class identity in the platform. Treating agents as a first-class identity means bringing them into the same system enterprises already use to manage people with visibility, access controls and governance built in. And the timing matters. Okta says 91% of organizations are already using AI agents in some form, but only 10% say they have a well-developed strategy to manage them. Let's explore how Okta for AI agents can close that gap. In the demo, watch how admins can register known agents into Okta's directory or surface shadow agents that weren't formerly onboarded. Employees can then access approved agents through a workflow that feels familiar inside Okta. Let Okta be the one that says, are you allowed to access it? What's the level of access you have? Okay, now let me go broker the exchange of access tokens for you. And by getting in the middle, we give our customer security teams that level of control and really the ability to kill that access if something goes wrong. Okta for AI agents became publicly available after a development cycle that unfolded alongside customers in real time. One example of this is really a new feature that we're rolling out next month called agent to agent connections. And this came directly from one of our beta customers. They're building a homegrown agent that they have onboarded to Okta for AI agents, the EA version of it. And they now want to connect that agent to other agents. And so we built out based on that feedback and rather quickly, agent to agent connections, meaning an AI agent can both be an identity principal and it can be a downstream resource that other agents can access. So yeah, the whole product development lifecycle is now AI enabled, which is really compressing how quickly we can iterate and innovate with customers. So as AI agents move deeper into the enterprise, the real challenge isn't just deploying them. It's knowing how to control them. In part three, we look at what it actually takes to scale a solution like this across a real enterprise environment. Signing off for Okta's newsroom, I'm Diana Blass.

TL;DR

Okta for AI agents treats autonomous AI systems as first-class identities within the existing Okta platform, bringing enterprise-grade access controls and governance to AI agent deployments.
While 91% of organizations use AI agents, only 10% have well-developed management strategies—Okta addresses this gap by enabling admins to register agents, surface shadow agents, and broker access through centralized workflows.
The platform now supports agent-to-agent connections, allowing AI agents to both authenticate as principals and serve as resources for other agents, developed through real-time customer collaboration during the beta phase.

Summary

This video introduces Okta's approach to managing AI agents as first-class identities within enterprise environments. As organizations rapidly adopt AI agents—with 91% already using them but only 10% having a developed management strategy—Okta for AI agents extends the company's existing identity platform to bring visibility, access controls, and governance to autonomous AI systems. The solution allows administrators to register known agents, surface shadow agents, and broker access through familiar workflows. A key innovation highlighted is agent-to-agent connections, developed in response to customer feedback, which enables AI agents to function both as identity principals and as downstream resources that other agents can access. The platform leverages Okta's centralized identity management to give security teams control over what agents can do, including the ability to revoke access when needed.

Chapters

0:00 - The AI Agent Control Challenge
0:43 - Introducing Okta for AI Agents
1:37 - Platform Demo and Capabilities
2:17 - Agent-to-Agent Connections Feature

Key Quotes

0:23 "It's almost like the only human control that our customers have over a process that might just run off and do its own thing is the identity access authorization controls that they can enforce at scale across their enterprise."
0:55 "In Okta for AI agents, you can really think about it as an extension of our existing Okta platform, but ensuring that AI agents are treated as a first-class identity in the platform."
1:51 "Let Okta be the one that says, are you allowed to access it? What's the level of access you have? Okay, now let me go broker the exchange of access tokens for you. And by getting in the middle, we give our customer security teams that level of control and really the ability to kill that access if something goes wrong."

FAQ

How does Okta for AI agents differ from traditional identity management?

Okta for AI agents extends the existing Okta platform to treat AI agents as first-class identities alongside human users, applying the same visibility, access controls, and governance frameworks that enterprises already use for workforce identity management. This includes registering agents in the directory, brokering access tokens, and providing security teams the ability to revoke agent access when needed.

What are agent-to-agent connections and why do they matter?

Agent-to-agent connections allow an AI agent to function both as an identity principal that authenticates to access resources and as a downstream resource that other agents can connect to. This capability, developed based on beta customer feedback, enables more complex agentic workflows where AI systems need to interact with each other while maintaining proper access controls and audit trails.

Categories:

» Data Protection

Tags:

Show more Show less

TL;DR

Summary

Chapters

Key Quotes

FAQ

Action1: The Remediation Gap: Vulnerability Management in the Age of AI

The AI-Powered VMware Alternative

LATAM: Accelerating Insights on AI Through an Engaging Webinar Series

Generative AI Security: Preventing AI from Becoming a Data Breach Multiplier

Schutz von KI in Anwendungen, Agenten und APIs.