Getting Started¶

Welcome to the AI Agent Builder Workshop!

In this hands-on workshop, you'll build Travel Assistant, an AI-powered agent that assists users plan trips, search flights, and provide personalized recommendations.

Foundation Labs¶

If you're new to the Microsoft Agent Framework, we recommend starting with the foundation labs in labs/00-foundations/.

These standalone labs introduce core concepts through hands-on examples. The code in these labs is separate from the main travel assistant codebase, so you can experiment freely.

Basic Agent - Create your first agent with multi-turn conversations
Context Provider - Add dynamic context to agent responses
RAG (Retrieval Augmented Generation) - Implement semantic search over documents
Long-Term Memory - Persist user preferences across sessions
Tools and Function Calling - Enable agents to perform actions and call external APIs
Middleware - Add PII filtering and cross-cutting concerns
Skills - Use file-based skills for progressive disclosure of capabilities
Hosting - Expose agents as web services with OpenAI-compatible endpoints

Travel Assistant¶

The codebase for the workshop is organized into two main parts: the backend (.NET) and the frontend (React).

src/
├── backend/    # .NET backend code
└── frontend/   # React frontend code

Technologies You'll Use¶

Microsoft Agent Framework - SDK for building intelligent, context-aware agents with built-in support for orchestration, memory management, and tool integration.
Azure AI Foundry - Used for accessing Azure OpenAI models for inference and generating embeddings.
Azure Cosmos DB - NoSQL database service used for storing agent memory and application data.
.NET/C# - Backend development of the agent's logic and API.
OpenTelemetry - A standard for observability, used for tracing and monitoring the agent's execution.
React - Frontend development for the user interface.

Let's Get Started¶

Head over to the Environment Setup page for instructions on setting up your workshop environment.

Happy coding!