AI agents are already in your clusters. Let’s talk about it in Salt Lake City.

Hi John,

AI agents are running in production clusters today. They call tools over MCP, call each other, and reach external model providers. Most arrived through application teams, not the platform team.

On Monday, November 9, during KubeCon + CloudNativeCon North America, Tigera CTO Peter Kelly walks through the path teams take to get those agents under control, in order:

• Discovery. Find the AI agents nobody registered.
• Identity. A service account tells you which pod made a call, not which agent, acting for whom, with what authority.
• Governance. Write policy at the level of tools and data, not IPs and ports, and run it in observation mode until you trust it.

The session is architecture-first and vendor-neutral. You’ll leave with a model you can apply to whatever you’re building on.

Monday, November 9 | 2:00–7:00 PM MST
Le Méridien Salt Lake City Downtown, five minutes from the Salt Palace

In-person seats are limited, so attendance is by invitation. Can’t make it to Salt Lake City? Stream it live.