Join Zenity and EverSec Group experts on March 12, 2026, for an exclusive NYC dinner featuring live nigiri carving, curated sushi pairings, and a practical discussion on the real-world challenges of securing AI agents.

As AI agents rapidly move from experimentation into production, security defenders are being asked to protect systems that are autonomous, adaptive, and deeply integrated into business workflows. This short educational presentation provides a practical overview of the emerging resources available to help defenders secure agentic AI systems across their lifecycle.

The session is structured around a familiar security process. It begins with the NIST Cybersecurity Framework (CSF) 2.0 Cyber AI Profile, establishing governance, risk management strategy, and organizational roles for the secure adoption of AI agents. From there, the focus narrows to implementation-specific risks using the OWASP Agentic Applications Top 10, highlighting the most critical threats and design pitfalls defenders should anticipate when building and deploying AI agents. Finally, the presentation connects these risks to real-world adversary behavior through MITRE ATLAS, demonstrating how known attack techniques against AI systems can be used to inform detection, response, and runtime protections.

Attendees will learn how these frameworks complement one another, how to apply them to threat modeling and control selection for AI agents, and how defenders can translate strategic guidance into operational security practices. The presentation is designed for security teams seeking a clear, end-to-end approach, and critically, a shared language for securing AI agents in production environments.