KnowBe4 debuts guardrails for autonomous AI agents

Security Brief Asia reports that KnowBe4 has expanded its human risk management portfolio into the autonomous agent arena with the launch of Agent Risk Manager, a real-time monitoring and governance layer engineered to police the actions of AI agents that are increasingly operating beyond the direct supervision of human employees.

The tool addresses a widening security gap created by the shift from simple AI assistance to autonomous systems capable of manipulating data and interacting with enterprise tools, targeting threats such as unauthorized sensitive data exposure, prompt injection jailbreaks, and runaway consumption of compute resources.

Greg Kras, KnowBe4's chief product officer, noted that "securing the prompt is only half the battle," emphasizing that the product focuses on the output and behavioral guardrails of these digital workers to prevent them from becoming sophisticated backdoors or uncontrolled shadow IT. The platform leverages over two dozen classifiers to redact credentials before they hit audit logs and maintains a dynamic inventory of agents and their permissions.