Broadcom Inc. (NASDAQ: AVGO) introduced AgentMinder, a new enterprise solution designed to manage and control autonomous artificial intelligence agents, at VMware Explore 2026 in Las Vegas.

According to a press release, AgentMinder functions as a governance layer that verifies agent identity and authorizes each action against a declared mission, intent, context, and current risk level before the action reaches an enterprise resource. The solution targets organizations deploying AI agents that access data, invoke tools, and execute business processes.

The product offers three core capabilities: identity and intent management, which binds an agent's authority to a declared mission and approved resources; runtime enforcement via a cloud-native AI gateway that evaluates context and enforces policy compliance; and an observability layer built on OpenTelemetry that provides audit trails, anomaly detection, and chain-of-custody tracking.

AgentMinder integrates with existing authorization systems through the AuthZEN standard, allowing organizations to reuse current policy enforcement endpoints. The solution is deployable on-premises, in virtual private clouds, or across public cloud environments, including VMware vSphere Kubernetes Service and Google Cloud Platform.

Broadcom states it uses AgentMinder internally, with the system handling approximately 36 million customer-related and seven million workforce-related API calls per day across more than 20 million customer identities and 72,000 workforce identities.

"With AgentMinder powering our platform, we've achieved massive global scale paired with zero downtime, even during maintenance and upgrades," said Alan Davidson, CIO of Broadcom.

Clayton Donley, vice president and general manager of Broadcom's Identity Management Security Division, noted that existing security models were designed for human users rather than autonomous agents, adding that AgentMinder provides verification and tracking for AI agent activity.