Building an AI agent is only the beginning. Connecting it to enterprise knowledge, enabling it to take action, and understanding how it behaves are essential steps toward production. Microsoft Foundry brings these capabilities together, but how do the pieces fit, and where should you start?

This 75-minute session provides a practical tour of Microsoft Foundry through the lifecycle of an AI agent. We’ll explore how Toolbox connects agents to tools and services, how Foundry IQ grounds responses in enterprise knowledge, and how Hosted Agents supports deploying and running custom agents. From there, we’ll examine how Observability helps teams trace execution and investigate issues, how evaluations measure quality and identify regressions, and how Guardrails help address safety risks and enforce controls.

A shared scenario will connect these capabilities, highlighting their roles, integration points, and the decisions teams face as they move from experimentation toward production. Attendees will leave with a clear understanding of the platform and a practical starting point for their own solutions.

You will learn:

• How Microsoft Foundry’s core capabilities work together across the agent lifecycle.
• To identify where tools, enterprise knowledge, and managed hosting fit into an agent solution.
• How observability, evaluations, and guardrails support quality, safety, and ongoing improvement.