What happened

Okta used its Oktane event to explain how AI agents will be governed when they interact with identity systems. The company introduced an Agent Gateway that moves the identity provider from simply granting access to also evaluating agents' actions at runtime. Charlotte Wylie, senior vice-president and deputy chief security officer at Okta, said the Blueprint Alliance will provide a reference architecture for securing the whole agentic stack. She added that the industry must “galvanize” to solve the problem quickly.

Ric Smith, president of products and technology at Okta, said the shift means Okta can now make more decisions about what it authorises. He told theCUBE, “That is a very different transition for Okta.”

Inside Okta, the Dex platform gives employees access to specialised agents without multiplying the interfaces they must navigate, according to Jenna Cline, senior vice-president of technology, data and intelligence at Okta. Dex can call a sub-agent that is scoped to a single task, making the approach secure and efficient.

Okta also highlighted partnerships with 1Password and ServiceNow. 1Password’s Credential Broker supplies secrets without exposing them to the agent, as explained by Nancy Wang, chief technology officer of 1Password. ServiceNow maps agents to business processes and is working with Okta on agent identities and session tokens, said Bhakti Pitre, vice president of AI platform security product at ServiceNow.

Why it matters

AI agents can call tools, delegate tasks and act across multiple applications. Without consistent controls, an agent could gain excessive privileges or act on outdated authorisation, exposing organisations to risk. Krista Case, principal analyst at theCUBE Research, warned that conflicting insights across platforms could leave gaps in authority. The new shared runtime controls aim to ensure that every action an agent attempts is evaluated against a central policy, reducing the chance of unauthorised activity.

The Blueprint Alliance’s reference architecture is intended to give customers a common point where controls can govern attempted actions. By moving into the runtime path, Okta’s Agent Gateway expands the identity provider’s role from granting access to evaluating what agents seek to do. This shift helps keep the agentic stack under continuous oversight rather than a single point-in-time check.

What this means for enterprise IT teams

IT teams will need to adopt a shared architecture that spans identity providers, secret-management tools and ticketing platforms. Integrating Okta’s Agent Gateway with existing identity solutions will allow teams to monitor and decide on agent actions in real time. Partnerships such as 1Password’s Credential Broker mean that agents can obtain secrets without exposing them, adding a layer of protection.

ServiceNow’s work on mapping agents to business processes will help teams intervene when an agent behaves unexpectedly. Teams should also consider how Dex can be used to scope agents to single tasks, reducing the attack surface. Coordinating across vendors will be essential to build a unified governance framework that can keep pace with the rapid adoption of AI agents.

In the words of Ric Smith, president of products and technology at Okta: "That is a very different transition for Okta."

The bottom line

Okta’s announcements at Oktane point to a future where shared runtime controls, industry alliances and cross-vendor standards protect against misuse of AI agents. Enterprises that adopt these measures early will keep their AI-driven processes secure while maintaining efficiency gains.

Sources

  • siliconangle.com