AI agent adoption outpaces enterprise governance

Okta reports AI agent adoption far outpacing governance, exposing identity control gaps across enterprises.

•91% of enterprises run AI agents with no governance framework in place

•The governance gap reveals a growing loss of control as autonomous agents proliferate across enterprises

The fact

Okta's Businesses at Work report, based on over 8,000 enterprise integrations in the Okta Integration Network, finds that 91% of organisations are already using AI agents while 90% lack a comprehensive governance strategy. Only 32% apply the same level of scrutiny to AI agents as to human users. The report also finds that 42% of organisations now operate with widespread non-human identities.

The Assessment

Enterprise adoption of AI agents is outpacing identity and access governance, creating a structural control gap between deployment and oversight. AI agents are increasingly acting as persistent, permissioned non-human actors within enterprise environments, while existing IAM frameworks remain human-centric. This shifts identity systems from user management tools into de facto control planes for autonomous execution, exposing organisations to visibility and auditability gaps rather than purely model-level risk.