Building a Security Blueprint for the Ungoverned Agent Problem

AI agents are no longer a future consideration. They are operating inside enterprise environments today, authenticating with legitimate credentials, inheriting permissions no one reviewed, and acting at a speed and scale that existing identity and access controls were never designed to address.

In this session, Mark Clancy, CISO, T-Mobile, Jason Lish, CISO, Cisco, and moderator Ryan Knisley, former CISO, Walt-Disney & Costco, share how their security organizations are responding to the "agentic era". They discuss initial governance missteps to the runtime enforcement models and risk-segmentation frameworks now in place.

KEY TAKEAWAYS

• Why conventional security tools have a structural blind spot for non-human identities and AI agent behavior
• How to segment AI risk across the enterprise and apply differentiated controls without impeding adoption
• What a governance model looks like when it's designed to accelerate the business
• The questions boards are asking about AI agent risk and what a credible, defensible answer requires