About This Book
Who this book is for
This book is for intermediate C# developers and software architects who are ready to move beyond AI prototypes and build production-ready intelligent applications. If you understand .NET fundamentals and want to integrate large language models into real business workflows using Microsoft's unified AI stack, this book provides the architectural guidance and practical implementation details you need.
What you will learn and build
You will construct Contoso Support Copilot, a comprehensive support assistant that evolves throughout the chapters. Starting with Microsoft.Extensions.AI as your core abstraction layer, you will add Semantic Kernel for plugin orchestration and prompt management, then implement retrieval-augmented generation using Azure AI Foundry and vector search. You will progress to designing autonomous agents with Microsoft Agent Framework 1.0 and coordinating multi-agent workflows for complex triage and escalation scenarios. Along the way, you will master structured output validation, defensive prompt engineering, state management, and AI-specific security boundaries. By the final chapter, you will have hardened your application with observability, automated testing strategies, and performance tuning suitable for live environments.
How to use this book
The book follows a cumulative approach where each chapter extends the Contoso Support Copilot solution. While you can reference individual chapters for specific techniques, sequential reading from Part I through Part V yields the best results. Parts I and II establish your foundation with model connectivity and service architecture. Part III adds knowledge retrieval capabilities, and Part IV introduces agent orchestration. Part V operationalizes everything for production deployment. Code samples and configuration guidance assume you are working through the solution incrementally.
High-level prerequisites
You should have intermediate proficiency in C# and familiarity with ASP.NET Core or similar .NET application frameworks. Basic understanding of Azure services, REST APIs, and dependency injection patterns is expected. No prior AI or machine learning expertise is required, though conceptual familiarity with large language models and embeddings will help you move through early chapters more quickly. You will need Visual Studio or VS Code, an Azure subscription for Azure AI Foundry access, and .NET 10 SDK.
