KnowBe4 today announced the launch of Agent Risk Manager, the industry’s first dedicated defense system designed to secure, monitor, and govern autonomous AI agents as they become embedded across enterprise workflows. The new solution expands KnowBe4’s Human Risk Management Plus (HRM+) platform, addressing what the company calls a rapidly emerging “agentic paradox”: empowering AI agents to act productively without exposing organizations to new, destructive risks.
As businesses transition from AI-assisted workflows to agent-managed operations, security leaders are facing a widening gap. Traditional approaches such as static code scanning and basic API security fail to address the real-world behavioral risks posed by autonomous agents once they are deployed. Agent Risk Manager is built specifically to close that gap by focusing not just on prompts, but on agent outputs, actions, and decision paths across the enterprise.
“We are moving from a world of human risk to universal risk—where both people and autonomous AI agents must be secured with equal rigor.”
— Greg Kras, Chief Product Officer, KnowBe4
“The industry has spent years securing the human element, but today, AI agents are effectively the newest members of the workforce,” said Greg Kras, chief product officer at KnowBe4. “Securing the prompt alone is not enough. Our Agent Risk Manager ensures that agents don’t become shadow IT or an unintentional backdoor for advanced prompt injection attacks.”
At its core, Agent Risk Manager provides a real-time operational control layer for AI agents. Key capabilities include behavioral guardrails that monitor agent actions to prevent unauthorized data exfiltration, agentic identity governance to define what tools and permissions agents can access, and adversarial simulations that stress-test agents against evolving hacker tactics. The solution is powered by more than 15 years of KnowBe4 behavioral data, enabling predictive detection when an agent deviates from safe operating parameters.
The platform also delivers prompt injection detection, sensitive information redaction, unbounded consumption monitoring to stop runaway agents, and automatic agent inventory across tenants. A compliance-ready audit log and centralized integration management further support rapid incident response and governance at scale.
Customer feedback underscores KnowBe4’s growing AI focus. A recent review on TrustRadius praised the company’s AI-driven personalization, noting its effectiveness in reducing user fatigue while delivering targeted, relevant security messaging.
