Most developers have already added an LLM to an application. The next challenge is building software where AI can reason, plan, use tools, and collaborate with your existing code. All while giving you full control over where models run.

Microsoft's open-source Agent Framework provides a flexible foundation for building AI-native applications. Whether you're using Azure OpenAI, OpenAI, Anthropic, or local models running with Ollama for privacy, compliance, or sovereign AI scenarios, Agent Framework offers a consistent programming model for orchestrating intelligent workflows.

In this demo-rich session, you'll learn how to evolve from simple prompt-based interactions to building your own agent harness capable of planning, invoking tools, maintaining memory, and executing multi-step tasks. We'll explore the core building blocks of Agent Framework, show how it integrates with existing .NET applications, and demonstrate how to combine cloud and local models depending on your requirements.

You'll leave with a clear understanding of when to use agents, how to architect AI-native applications, and how Microsoft Agent Framework helps you build production-ready solutions that go far beyond chatbots.

You will learn:

• The architecture of AI-native applications
• How to build intelligent agents with Microsoft Agent Framework
• How to integrate cloud or local LLMs into your .NET applications