Oktane 2026: Setting the standard for the secure agentic future

AI agents promise genius, but only if organizations can deploy them securely, says Okta CEO.  

About the Author

23 septiembre 2026 Time to read: ~

Vidyard video

The promise of AI agents goes beyond automation. The real prize is agentic genius—creating products no one has imagined and business insights no one knew to look for. But to get there, organizations face a trade-off: restrict AI agents and lose their value, or grant them full freedom and risk exposure.

At Oktane 2026, Okta CEO and Co-Founder Todd McKinnon addressed this dilemma in his keynote, advocating for more visibility and control to deliver transformational value without sacrificing security.

"Agents need deep, broad, persistent access to systems to deliver value," McKinnon told the audience. "The challenge is doing that securely, at scale. We must eliminate the tradeoff between security and innovation."

Okta's answer comes in three parts, all announced at Oktane:

The blueprint for the secure agentic enterprise: This new reference architecture gives organizations a standardized way to answer four critical questions: Where are my agents? What can they do? What are they doing? How do I respond? The blueprint was developed by the Blueprint Alliance, a coalition of industry leaders—including AWS, CrowdStrike, Databricks, Google Cloud, Salesforce, and ServiceNow.

Agent SSO: Built on Cross App Access, an open protocol that enables secure agent-to-app and app-to-app connections, Agent SSO allows enterprises to control how AI agents connect with applications. With Agent SSO, organizations have a common way to find agents, define and understand what they can do, and respond when something goes wrong.

Okta for AI Agents: Developed in collaboration with customers, Okta for AI Agents gives enterprises a practical onramp for putting the blueprint into action. With new capabilities—including Shadow AI Agent Discovery for Endpoints, Agent Gateway, and third-party identity provider support—Okta for AI Agents allows organizations to manage agents as first-class identities. 

While these innovations are important, no single vendor can secure the agentic enterprise alone. Because AI agents traverse models, data, and infrastructure, the entire industry must share signals and enforce unified controls as one cohesive system.

"The secure agentic enterprise isn't some distant future," McKinnon said. "It's being built now. But it requires all of us working together."

Watch the full keynote above to learn how Okta and other enterprises are bringing the blueprint to life. 

About the Author

Get our Identity newsletter

Access Granted Newsletter