IBM Launches New Security and Identity Features for Enterprise AI Agents

IBM announced new capabilities to manage and secure AI agents. These features debuted at the TechXchange 2026 conference. They focus on identity verification, access control, and governance for AI agents operating across business systems.

The announcement features the general availability of "IBM Agent Identity." It also includes an integration with HashiCorp Vault. This combination enables enterprises to create a central registry for AI agents. It assigns clear ownership and manages permissions with short-lived credentials. This enforces "least-privilege" access. The system creates a traceable audit trail for all agent actions.

These tools prevent "rogue agents" from disrupting operations or creating unexpected costs. The capabilities integrate with other IBM products, including watsonx Orchestrate. They also integrate with partner technologies, such as NVIDIA's OpenShell. This provides a comprehensive security architecture for AI.