Microsoft has used its Azure platform to launch its OpenAI efforts. Recently, it rebranded and renamed its platform to Microsoft Foundry, presumably with some staying power. Like all new technical applications, Foundry can be tedious to use, and its online documentation is sprawling. Eduardo Sojo attempts to make sense of this application and explain it to new users so that they can make the most of it in their personal and professional applications.
Much of this book is simply technical writing to walk through the features of Microsoft Foundry. It describes, step by step, the configurable options to deploy an AI model and their meaning. The depth surpasses what you can gather from the interface, but it lacks the depth of AI theory.
The interesting part takes flight in the last three chapters where Sojo describes how to build AI agents and their implications. He uses an extended example from the finance industry, where data security is at a premium. He ends with a two-agent model that involves cross-communication via Model Context Protocol, or MCP.
Like all technical manuals, this book will serve its purpose for a time and then pass on. It helped me meet my goal of understanding the Foundry in more depth so that I can exploit its features better. It provides a necessary missing piece that helps users configure and deploy AI to meet their goals.
Microsoft Foundry in Action: A Hands-On Guide to Building, Evaluating, and Governing Enterprise AI Agents
By Eduardo Sojo
Copyright (c) 2026
Packt Publishing
ISBN13 9781835888698
Page Count: 526
Genre: Artificial Intelligence
www.amazon.com