Agentic AI with Microsoft Foundry: Design and develop intelligent AI solutions and autonomous agents with Microsoft's Agent Framework

Master building purpose-built generative and agentic AI solutions with Microsoft Foundry and the Agent Framework, integrate enterprise data, fine-tune OpenAI models, and deploy secure, production-grade copilots on Azure.

Key Features

• Build copilots and autonomous agents using Microsoft Foundry
• Fine-tune OpenAI models and integrate enterprise knowledge
• Deploy secure, scalable, and responsible AI solutions on Azure
• Purchase of the print or Kindle book includes a free PDF eBook

Book Description

Agentic AI with Microsoft Foundry is your complete guide to creating intelligent, enterprise-ready AI copilots and agents using Microsoft’s unified AI development platform. Whether you’re building with GPT models, integrating private data, or orchestrating multi-agent workflows, this book equips you with the technical foundation and practical skills to succeed.

You’ll begin by mastering Microsoft Foundry essentials, including setting up your workspace, exploring the Model Catalog, and applying prompt engineering techniques for high-quality LLM outputs. You’ll then fine-tune large language models, implement retrieval-augmented generation (RAG), and integrate cognitive search capabilities to give your AI real-world context.

The second half of the book dives deep into building and extending AI agents using the Agent Framework, covering everything from configuring tools and connectors to orchestrating multi-agent systems capable of reasoning, retrieving, and acting autonomously. You’ll also learn how to evaluate and govern your AI responsibly, monitor deployments effectively, and scale solutions for enterprise use.

By the end, you’ll have built a production-ready AI copilot that leverages Microsoft Foundry, OpenAI models, and Microsoft’s Agent Framework, bringing together intelligence, automation, and ethical AI design.

What you will learn

• Configure and navigate Microsoft Foundry projects
• Build copilots using GPT and OpenAI models
• Apply prompt engineering and fine-tuning techniques
• Implement retrieval-augmented generation (RAG)
• Create intelligent agents with the Agent Framework
• Connect enterprise data and external APIs securely
• Orchestrate multi-agent workflows and automation
• Evaluate, monitor, and deploy responsible AI systems

Who this book is for

For AI developers, data scientists, cloud engineers, and solution architects looking to design, fine-tune, and deploy intelligent AI solutions using Microsoft Foundry and the Agent Framework. Suitable for professionals exploring generative AI, Microsoft Copilot integrations, or autonomous enterprise agents. Basic familiarity with Python or cloud services is recommended but not required.

Table of Contents

1. Introduction to Generative and Agentic AI
2. Setting up Your Microsoft Foundry Environment
3. Core Concepts and Tools of Microsoft Foundry
4. Exploring Model Catalog and Evaluation
5. Fine-Tuning Models for Custom Solutions
6. Building Your First Agent with Foundry
7. Integrating Enterprise Data for Contextual AI
8. Advanced Agent Capabilities and Tool Integration
9. Multi-Agent Orchestration and Workflows
10. Evaluating AI Solutions and Responsible AI
11. Deploying and Integrating AI Applications
12. Scaling Solutions and Future Trends

Balamurugan Balakreshnan is a principal cloud solution architect at Microsoft Data/AI Architect and Data Science. He has provided leadership on digital transformations with AI and cloud-based digital solutions. He has also provided leadership in terms of ML, the IoT, big data, and advanced analytical solutions.

Sina Fakhraee, Ph.D., is currently working at Microsoft as an enterprise data scientist and senior cloud solution architect. He has helped customers to successfully migrate to Azure by providing best practices around data and AI architectural design and by helping them implement AI/ML solutions on Azure. Prior to working at Microsoft, Sina worked at Ford Motor Company as a product owner for Ford's AI/ML platform. Sina holds a Ph.D. degree in computer science and engineering from Wayne State University and prior to joining the industry, he taught various undergrad and grad computer science courses part time.

Jay Padhya is a Senior Cloud Solution Architect at Microsoft with a passion for solving complex problems and helping enterprises achieve more with technology. He has extensive experience in data science, enterprise architecture, and scalable solution delivery across industries. Before Microsoft, Jay was a Senior Data Scientist at CVS Health-Aetna and Lead Data Scientist at Stellantis, where he built and optimized machine learning solutions in healthcare and automotive domains. His earlier roles include data analysis and visualization at Exide Technologies, and serving as a Business Analyst leading technical implementations for SaaS platforms. Jay holds a Master's degree from Northeastern University and is a certified Business Analyst (IIBA, Canada). He also completed a co-op as a FIX Data Scientist at Charles River Development, contributing to financial algorithm development and cloud deployments.

Minsoo Thigpen is a Principal Product Manager at Microsoft, where she leads product initiatives for generative AI safety evaluations and automated red teaming within Azure AI Foundry, Microsoft's platform for building and securing generative AI systems. She develops enterprise-grade tools that help organizations measure model safety, uncover vulnerabilities, and ensure the security of AI models, applications, and agentic systems. With more than seven years in the Responsible AI space, Minsoo has helped shape how practitioners assess and govern AI through both open-source contributions and engineering leadership. Her work includes contributions to interpretability, fairness, and accountability toolkits such as InterpretML, Fairlearn, Error Analysis, and the Responsible AI Toolbox. At Microsoft, she has built scalable, automated safety evaluation and AI red teaming frameworks that support proactive risk identification and mitigation in generative AI workflows. Minsoo's contributions span technical innovation and community impact, empowering developers and enterprises to adopt trustworthy AI practices at scale.