AI Agent Rollouts Push Enterprises to Overhaul Fragmented Identity Stacks

Enterprise deployments of artificial intelligence agents are forcing corporate security teams to confront decades of fragmented identity and access management infrastructure, as the pressure to deploy autonomous software exposes integration costs and security gaps across legacy tools.

Speaking at Okta Inc.’s Oktane event in an interview broadcast by SiliconANGLE Media’s livestreaming studio theCUBE, Okta Chief Revenue Officer Jon Addison explained that organizations are treating agentic security not as an entirely new discipline, but as a direct extension of existing identity management frameworks.

Addison noted that many corporate networks rely on disparate point tools stitched together over years. Maintaining interoperability across those point solutions creates ongoing expense while introducing vulnerabilities that become more dangerous when automated software agents access network resources. Consequently, enterprises are working to consolidate architectures and establish foundational access controls before rolling out agents at scale.

Rollout readiness varies significantly across organizations. While some businesses are establishing formal reference architectures to govern permissions prior to deployment, Addison said others already have unauthorized "shadow agents" operating in production environments, creating an urgent need for retroactive access management controls.

Current enterprise adoption remains primarily focused on augmenting human employees and improving task efficiency. However, Addison warned that security teams must prepare identity architectures for an eventual surge in fully autonomous agents operating across corporate systems without direct human intervention.