Building AI Agents in .Net: A Code-First Approach Using Microsoft Agent Framework

By: Costea, Daniel

Get the eBook free when you register your print book at Manning.

AI Agents are rapidly moving from side projects to an integral part of modern application architecture. This book teaches you to develop effective, reliable, and enterprise-ready AI agents using Microsoft Agent Framework. You'll learn how to design and deliver agents that are testable, observable, and deployable inside existing systems, scaling from local prototypes to production. Working with hands-on code, you'll define agents, attach tools and memory providers, and orchestrate interactions, all integrated with dependency injection, logging, and middleware.

You'll work to assemble Robby, a robot car assistant, evolving it from a minimal agent, then deepening our understanding of chat clients and model providers. From there, you'll build complete agents from scratch, extend them with tools, memory, and Model Context Protocol, and harden them with middleware, observability, and evaluation. Finally, you'll put Robby into multi-agent scenarios using sequential, concurrent, collaborative, and hierarchical patterns, finishing with custom workflows for long-running, stateful journeys.

What's Inside

- Integrate agents into your .NET applications
- Work with chat clients for proprietary, open-source, and self-hosted models
- Connect to external systems via tools and Model Context Protocol
- Add memory for longer conversations and processes
- Apply middleware, observability, and evaluation for production-grade agents (for example: logging, validation, auditing, and measurement)
- Orchestrate multiple agents using high-level orchestration patterns and low-level custom workflows

About the Reader

This book is for .NET developers, software architects, and technical leads building practical agentic AI systems in C# with Microsoft.Extensions.AI (MEAI) and Microsoft Agent Framework

About the Author

Daniel Costea is a software engineer with more than 20 years of experience in enterprise systems, product development, and developer training. He has been a Microsoft Most Valuable Professional (MVP) since 2019, recognized by Microsoft for his contributions in AI and .NET technologies. Daniel speaks frequently at technology conferences, has trained hundreds of developers, and created multiple courses on the practical, production-ready use of the modern .NET ecosystem.

• Format: Paperback
• Language: English
• ISBN: 1633435563
• ISBN13: 9781633435568
• Release Date: October 2026
• Publisher: Manning Publications
• Length: 450 Pages
• Weight: 1.18 lbs.