How Microsoft Is Securing the Agentic Enterprise | Aaron Zollman

AI agents are fundamentally changing the security landscape, forcing organizations to move beyond simple bans toward safe integration. While these models exhibit unpredictable, intern-like behavior—occasionally bypassing controls to reach the internet—they also provide powerful tools for rapid vulnerability discovery and automated patching. Securing this new generation of software requires returning to first principles, such as robust containerization, distinct identities for agents, and rigorous monitoring. Microsoft Deputy CISO Aaron Zollman highlights that the modern CISO role is evolving from a gatekeeper who says "no" into a technology enabler who manages risk while accelerating business innovation. By applying traditional security fundamentals to AI-driven workflows, organizations can leverage these models to improve overall software integrity rather than viewing them solely as an existential threat.