AI agents can look like magic, but running them in production is an operational challenge. Every action depends on Azure services that must be secured, monitored, governed, and kept under control.

In this demo-driven session, we’ll follow real agent activity across Azure OpenAI, Search, Foundry, and other supporting services. You’ll see what happens when an agent reasons, retrieves data, calls tools, and connects to enterprise systems.

We’ll cover identity boundaries, safety controls, observability, troubleshooting, performance, and cost management. If you want a practical, operator-focused guide to running AI agents in production, this session is for you.