Traditional identity management platforms rely on periodic manual reviews, a model struggling to keep pace with the rapid proliferation of non-human identities and AI agents. Lumos addresses this imbalance by deploying specialized agents—including an Access Review Agent and an NHI Owner Hunter—that operate in the background to handle granular permission decisions. By mapping every identity and permission in real-time, these agents identify risks and enforce least-privilege access without requiring constant human intervention.

According to Andrej Safundzic, CEO of Lumos, the current security landscape forces enterprises into an untenable choice between speed and safety. The new Agent Hub provides a way to fight back by offloading repetitive tasks to machine-speed workflows. While the agents handle continuous monitoring and access revocation, human teams retain control by defining the overarching strategy and policy. This approach allows organizations like Mars, Netskope, and GitLab to maintain compliance and security in an era where the sheer volume of identities has outpaced human oversight.