There are plenty of questions to answer about The AI Playbook. Readers of this book come with diverse backgrounds and various preconceptions about the problem the book aims to solve: getting machine learning deployed. This FAQ will orient you, clarifying why you should read this book and aligning your expectations.
Why did you write The AI Playbook?
Outside Big Tech and a handful of other leading companies, machine learning initiatives routinely fail to deploy, never realizing value. What’s missing? A specialized business practice suitable for wide adoption. In The AI Playbook, I present the gold-standard, six-step practice for ushering machine learning projects from conception to deployment. I illustrate the practice with stories of success and of failure, including revealing case studies from UPS, FICO, and prominent dot-coms. This disciplined approach serves both sides: It empowers business professionals and it establishes a sorely needed strategic framework for data professionals.
What is this book about?
This book presents a strategic and tactical playbook for launching machine learning, a six-step discipline to run an ML project so that it successfully deploys. I call this practice bizML.
Along the way, the book also delivers the semi-technical background knowledge everyone participating in the project needs—in a friendly, accessible way anyone can understand. Because of that coverage, the book also serves as a non-technical introduction to the field for newcomers.
Why does machine learning need a specialized business practice?
Here’s the problem. ML is the world’s most powerful generally applicable technology. But ML can only improve large-scale operations by changing them. For that reason, an ML project shouldn’t be viewed as “a technology project.” Instead, to make an impact, it must be reframed as a business project meant to improve operational performance, with ML as only one component—one that’s necessary but not sufficient.
With the attention overwhelmingly focused on the technical portion and its execution, the industry has failed to establish a widely adopted business practice for executing the whole other half of a successful ML project. As a result, new ML initiatives routinely fail to deploy.
Who is this book for?
This book serves anyone who wishes to gain value with ML by participating in its business deployment, no matter whether you’ll play a role on the business side or the technical side.
First and foremost, I wrote this book for business professionals—the people who run the ML project, hold stakes in it, make decisions about it, or manage the operations that will be changed (and improved) by it. This includes executives, directors, managers, consultants, and leaders of all kinds.
But this book is for techies, too. If you’re a data scientist, ML engineer, or any kind of technical practitioner involved with ML, this book invites you to step back from the hands-on execution and gain a new perspective on the holistic paradigm within which you are contributing.
Is this book a how-to?
This book is a business how-to, but not a technical how-to. Unlike most ML books, it tackles the business practice instead of the technical practice. It presents a six-step business practice, bizML, for running an ML project.
This book does not delve deeply enough to guide data professionals in the technical how-to. That’s what the vast majority of other ML books are for. The ML methods they cover are only one ingredient. They constitute a key technical component of the project, but that component makes for only one of the six project steps covered in this book. Accordingly, one and only one chapter of this book, chapter 5, delves into core ML methods—it provides an accessible “crash course.”
This book also differs from most business books on ML, which present a strategic industry overview. Such books typically cover the topic from a higher level, without providing how-to guidance and without concretely detailing how ML integrates to deliver operational improvements.
I’m a technically trained data professional—why do I need this book?
This book establishes a sorely needed strategic framework, providing complementary business-side know-how that all great data professionals need to master. The real “data science unicorn” isn’t the person who knows every analytical technique and technology; rather, it’s the one who has expanded their skillset to also participate in a company-wide, business-oriented effort that gets their models deployed. After all, the soft skills are often the hard ones.
In so doing, this book does cover certain technical steps generally omitted by courses and books meant for data professionals, including how to fully establish the dependent variable (called the output variable in this book), how to prepare the data, and how to establish the performance metric (including why accuracy and a popular technical metric called AUC are usually the wrong choice)—all so that these choices align with business objectives and operational considerations.
On the other hand, know that this broadly accessible book is not the technical fare to which you’re likely accustomed. For some experienced data professionals, the best use of this book may be to give it a good skim—slowing down to give chapter 0 on the need for a specialized business-side practice and chapter 3 on evaluation metrics a complete read—and then passing it on to your boss or a key colleague.
What kind of AI does this book cover?
The buzzword AI can mean many things, but this book is about ML, which is a central basis for—and what many mean by—AI. To be specific, this book covers the most vital use cases of machine learning, those designed to improve a wide range of business operations. This book does not cover other areas that are also sometimes referred to as AI, including artificial general intelligence (hypothetical systems that would be capable of any intellectual task humans can do), natural language processing, rule-based systems, and computer vision.
Does this book pertain to generative AI?
Yes. Generative AI dazzles the world by writing text and producing images—but when it comes to improving operational efficiencies, classical ML (aka predictive AI) has long reigned supreme. However, generative AI is also well suited and stands to potentially beat out classical ML in some arenas. The bizML practice presented by this book also serves generative AI—for projects that apply generative AI to measurably improve great numbers of operational decisions. For either kind of technology, bizML gets you there, guiding the project to a successful deployment.
Does this book pertain to deep learning?
Yes. Although deep learning is more technically complex than many classical ML methods and tends to be applied for different classes of problems (more on image processing, for example, and less on customer prediction), the ML project discipline presented in this book applies and is equally needed. The organizational challenges of deployment are largely the same, no matter how the model being deployed operates on the inside.
Does this book pertain to predictive analytics?
Yes—predictive analytics is a major subset of ML. It is the application of ML methods for certain business problems. Alternatively, in many contexts, predictive analytics is simply a synonym for machine learning.
This article is excerpted from the book, The AI Playbook: Mastering the Rare Art of Machine Learning Deployment, reprinted with permission from the publisher, MIT Press. It is a product of the author’s work while he held a one-year position as the Bodily Bicentennial Professor in Analytics at the UVA Darden School of Business
Eric Siegel, Ph.D. is a leading consultant and former Columbia University professor who helps companies deploy machine learning. He is the founder of the long-running Machine Learning Week conference series, the instructor of the acclaimed online course “Machine Learning Leadership and Practice – End-to-End Mastery,” executive editor of The Machine Learning Times, and a frequent keynote speaker. He wrote the bestselling Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, which has been used in courses at hundreds of universities, as well as The AI Playbook: Mastering the Rare Art of Machine Learning Deployment. Eric’s interdisciplinary work bridges the stubborn technology/business gap. At Columbia, he won the Distinguished Faculty award when teaching the graduate computer science courses in ML and AI. Later, he served as a business school professor at UVA Darden. Eric also publishes op-eds on analytics and social justice.