Broadcom introduced AgentMinder, a new solution for governing autonomous AI agents, at VMware Explore 2026. The tool acts like a traffic controller. It verifies each agent’s identity and checks every action against its declared mission, intent, context and current risk before the action reaches an enterprise resource. Existing security models were built for human users, not for digital employees that decide and act on their own. AgentMinder treats agents as enterprise-grade identities. It uses a cloud-native gateway for runtime enforcement and applies dynamic policies to keep agents within approved tools and resources. The system also provides continuous authorization, full telemetry and audit trails so every step is observable and attributable. Capabilities include scoped delegation so child agents cannot exceed a parent’s boundaries. Broadcom positions the product as a way for companies to scale agentic AI while keeping security and compliance intact.
Broadcom introduced AgentMinder, a new solution for governing autonomous AI agents, at VMware Explore 2026. The tool acts like a traffic controller. It verifies each agent’s identity and checks every action against its declared mission, intent, context and current risk before the action reaches an enterprise resource. Existing security models were built for human users, not for digital employees that decide and act on their own. AgentMinder treats agents as enterprise-grade identities. It uses a cloud-native gateway for runtime enforcement and applies dynamic policies to keep agents within approved tools and resources. The system also provides continuous authorization, full telemetry and audit trails so every step is observable and attributable. Capabilities include scoped delegation so child agents cannot exceed a parent’s boundaries. Broadcom positions the product as a way for companies to scale agentic AI while keeping security and compliance intact.