A few weeks ago, Microsoft released its new agent framework for AI developers. The aptly named Microsoft Agent Framework is an open-source SDK and runtime designed to let developers build, deploy, and manage sophisticated multi-agent systems. It unifies the enterprise-ready foundations of Semantic Kernel with the innovative orchestration of AutoGen, so teams no longer have to choose between experimentation and production.

How does the new framework actually work in practice and should you use it for every new project going forward? That’s what I’ll discuss in this session. After a brief introduction, I’ll dive into code and explore
– how you can use it to build agents,
– how you provide our agents with tools,
– how you orchestrate multi-agent systems,
– and how you can use telemetry to trace the actions your agents are taking.

I will walk you through some Python code samples and test them live (C# samples are also available). In the end, we will conclude on the new framework’s capabilities and how you can best utilize it in your customer projects.