Enterprise Vendors Rush to Govern Uncontrollable AI Agents

Four enterprise software vendors have recently released standalone AI agent governance products in a span of just two weeks. Okta, IBM, Broadcom, and Dataiku launched their respective tools between August 24 and early September. These products approach the problem from different infrastructure layers: identity, orchestration, security, and observability.

The Okta Agent SSO, launched on August 24, focuses on replacing static API keys with short-lived, identity-governed tokens. Ric Smith, Okta's president of products and technology, framed it as an infrastructure necessity, stating that AI agents are becoming a primary interface for work but should not compromise security or visibility.

IBM's watsonx Orchestrate shipped its AgentOps Agent on August 31. This tool approaches governance from the orchestration layer, discovering agents built on Amazon Bedrock and importing them into a single control plane. A Trace Inspector allows teams to follow an agent's execution path step by step, while a custom evaluation system lets companies define what 'good' means for their own terms.

Broadcom unveiled AgentMinder at VMware Explore in Las Vegas, taking the security and runtime angle. This platform acts as a traffic controller for autonomous agents, verifying each agent's identity and intended mission before authorizing actions against enterprise resources.

Dataiku's Agent Management product rounds out the sequence, launched in September with general availability planned for October. It connects to various platforms, measuring agents against business key performance indicators rather than uptime.