The Decision Intelligence Handbook

Preface

The past few decades have brought astonishing improvements in data science, business intelligence, and artificial intelligence (AI). In response, organizations are increasingly determined to weave these technologies into the fabric of their everyday decisions. Indeed, a 2019 survey conducted by global consulting firm McKinsey found that better decision making can benefit a typical Fortune 500 company by as much as $250 million per year.

This reflects a big opportunity to improve organizational outcomes, even as it reflects the dismal state of organizational decision making today.

But we’re far from achieving this nirvana. In Fortune magazine, Alan Murray and Jackson Fordyce write, “Business leaders are so overwhelmed with data they’re struggling to function.” And today, many “data-driven” and “evidence-based” initiatives are falling short. The reason is, simply, that decision making is not really about data: it’s about achieving an organization’s outcomes, with data as a key ingredient, but still secondary to business outcomes. This incorrect focus on the data itself leads to data and AI work that isn’t well aligned with many organizations’ outcomes and desired goals.

Smart organizations are moving, instead, to “outcome-driven” decision making, with data and technology working “under the hood” to supercharge their choices.

Along the way, a new discipline has emerged to help them, called decision intelligence (DI). DI brings AI (including generative AI technologies like ChatGPT), data, human expertise, research, and more into an integrated framework that answers two questions: “If I take this action today, in this context, what will be the outcome?” and “What is the best action to take today to maximize the likelihood that I’ll reach my goals?”

DI is about ensuring that decision makers can use the most powerful technologies, and that decision-making systems present information in a way that feels natural and intuitive. DI moves organizations beyond simply using historical data, which provides information and insights about the present and past, to answering questions about the future.

Hi, we’re the authors of this book, N. E. Malcolm and L. Y. Pratt. Pratt co-invented DI (with Mark Zangari) in 2010. Since then, working with our team, we’ve helped DI to grow into a field so promising that Forbes asks whether it’s “the new AI.” The Gartner Group predicts that more than a third of large organizations will be using DI by the time this book is published; market-research firm MarketsandMarkets projects that DI will grow to a $22 billion market by 2027; and Chinese behemoth Alibaba names DI second on its list of top technology trends for 2023 (just after generative AI).

One of us—coauthor Pratt—wrote the first book on DI: Link: How Decision Intelligence Connects Data, Actions, and Outcomes for a Better World.¹ We’ve now built and delivered dozens of DI solutions for large and small commercial organizations, startups, and the public sector. DI projects—ours and those of other DI practitioners—have saved and generated many hundreds of millions of dollars for organizations worldwide, in addition to social and other nonfinancial benefits.

About This Book

This book is a practical, “roll up your sleeves” guide to how you can do DI, within your own organization or as a consultant or Decision Intelligence Service Provider (DISP) or Decision Intelligence Infrastructure Provider (DIIP) for others. It’s organized around a collection of nine DI “best practice” processes. We’ll walk you through each one, starting with how to decide if DI is right for your situation. We’ll show you how to go about designing a decision. By the time you work through the book, you’ll have a continuously improvable decision asset that is connected to data, AI, and more in a way that will drive competitive differentiation and success through better decision making.

But before we dive into the gnarly details, we feel it’s important for you to understand that you can start doing DI today. Seriously, we’re talking about 20 minutes from now: the time that it takes to get to the section called “Build Your First CDD, Right Now!” in the next chapter.

Decision Intelligence in a Nutshell

Simply put, DI helps organizations make better decisions. It helps decision makers understand how the potential actions they can take today (the things they can do) could affect their desired outcomes (the things they want to accomplish). To get from actions to outcomes, DI centers around a drawing called a causal decision diagram (CDD), which acts as a “decision blueprint.” The CDD lets you design a decision. Its purpose is to get everyone on the same page—technologists, decision makers, and even the stakeholders affected by the decision. To give you an idea of what a CDD looks like, Figure P-1 shows a very simple one.

Figure P-1. A simple causal decision diagram (CDD).

We’ll have more to say about CDDs in later chapters, but you can see a few things right off the bat. We draw actions on the left-hand side of the diagram and outcomes on the right. Between the two is a chain of consequences. (Note that these are consequences, not tasks, and are—as a rule—outside of your control after you take an action.)

Figure P-2 shows a more complex CDD, including some annotation showing where technology fits into a decision. (To look at the details, you can download a PDF from the book’s supplemental materials repository.)

Figure P-2. A CDD that helps building managers decide “When and how should I go about opening my building and keeping people safe from the pandemic?”

To draw a CDD, you document your desired outcomes and the actions you can take to achieve them. You capture these in a diagram that also shows your understanding of the cause-and-effect chains that connect actions to outcomes. Later, if you want, you can add data, evidence, models, analytics, and human expertise that inform your decision, so that your diagram isn’t just on paper but also can be simulated by a computer. In this form, the simulation based on the CDD is effectively a “decision digital twin,” providing an evidence-based way to determine which actions will be most effective in achieving your outcomes. We’ll walk you through the process for creating CDDs in this book.

At its most fundamental, DI is an integration and design discipline, connecting technologies together with each other and with human decision makers. And the CDD shows how these pieces fit together. Simply put, DI helps organizations leverage the best of human expertise, hand in hand with all sorts of technology. DI begins and ends with the group or person making the decision. Technology is secondary, used in support of better decisions. Not fully automated, yet not just based on human judgment, DI is a “hybrid” AI method—one of the fastest-growing technology markets.

The DI methodology holds that structured decision making can be represented as a set of well-defined processes. These follow a lifecycle, beginning with formulating the decision at hand and ending with retrospectively analyzing the effectiveness of the chosen course of action, and possibly reusing the decision or elements of it for future decisions.

That might sound like a lot, but an important aspect of DI is that it’s easy to do, especially at the start. Rather than asking you to think about decisions in a new way, DI simply asks that you document the way that you think about decisions today. You’ll find that just drawing a picture of a decision—as we’ll teach you to do in the next chapter—goes a long way.

Along the way, you’ll find that DI includes these important elements:

• Clearly defining decision requirements

• Representing decision making as a set of well-defined processes that follow a lifecycle, from formulating the decision to retrospectively analyzing its effectiveness and potential for reuse

• Following an iterative design process that incorporates data, analytics, and expert judgment; allows for multiple scenarios; and models different potential worlds

• Creating a CDD as a unifying graphical representation for a designed decision

• Integrating decision assets like data, human knowledge, and machine learning (ML) and AI models with elements of the CDD; this lets decisions be driven by data and more

• Emphasizing quality assurance and security

• Transforming into a decision-centric organization using organizational and cultural best practices

LLMs, OMG

As this book goes to press, ChatGPT and other large language models (LLMs) are turning the technology world upside down, supercharging writers, coders, scammers, and more. And DI is no exception. We are already seeing LLM technology provide incredibly valuable advice to our DISP clients, acting as a superpowered new collaborator in multiple phases of the DI processes you’ll read about here. In particular, we’ve seen LLMs surface actions, externals, outcomes, and unintended consequences that no one had previously considered, helping to reduce “tunnel vision.”

We see LLMs’ role in DI as a sort of “super Google,” giving decision modelers and decision makers much easier access to a wide range of assets that they can use to inform their decision making. But LLMs don’t do action-to-outcome simulations, so they’re complementary to the decision-reasoning methods we describe here.

Who Is This Book For?

This book is for you if you’d like to learn how to introduce DI to your organization or to your clients. You might be an executive who takes decisions seriously, combining the best of diverse human and computer knowledge to drive competitive advantage. You might be passionate about addressing climate change, but you know that there’s a lot of earth observation (EO) data that’s going unused because data scientists don’t know how to connect it to decision making. You might be a data or AI consultant or an employee in one of the emerging DISP companies, looking to differentiate your practice by providing something new and valuable. You might be an ML expert who wants to maximize the value of this important technology, or a head of analytics or business intelligence who needs a way to communicate with your internal clients so that your technology helps them with better evidence-based decisions.

We’ve written this book for the “insurgent” bottom-up perspective, as well as for the lucky few who have obtained centralized executive sponsorship to take DI organization-wide. Indeed, we wrote this book in collaboration with a G20 central bank in the process of doing just that, and the bank has adapted this book for its internal use.

What You Will Learn

After completing this book, you will:

• Understand the kinds of decisions organizations make and which ones DI can help with

• Understand how to create, read, use, maintain, and reuse CDDs

• Have a “starter kit” of DI documents and templates that you can tailor for your organization

And you’ll understand how to use DI to:

• Structure decision conversations around desired outcomes (financial or not) and actions to achieve them

• Use state-of-the-art collaborative tools to map AI, knowledge, data, and more into decisions

• Find simplicity and order amid the confusion of complex data, tools, and decisions

• Provide value to your data management projects by guiding them toward the 10% of data that has 90% of the value

• Identify ways to integrate silos within your organization

• Specify your requirements for automated decision-reasoning simulations to a software team

• Earn greater trust and credibility for your data/analytics/AI team, because you speak your customer’s language

Please note that there are several topics that are not covered here. For instance, we don’t delve into the broader societal impacts of DI or its potential for solving complex problems like the climate and pandemics. These impacts are covered in Link. Finally, we don’t get into the technical specifics of how to build DI tools such APIs, interfaces, or AI and statistical models that interoperate with computerized DI models. Those technologies change quickly, but the principles we offer here stand on their own, independent of specific technological choices.

Assumptions This Book Makes

We do not assume that you have any specific technical knowledge. We have written this book to help all the participants in a decision-making process—not only the executives, managers, and stakeholders, but the analysts and data scientists who provide data and other evidence to decision makers.

We also do not assume that you have read Link. We’ll introduce everything about DI that you need to know. Where Link was a visionary survey of the field, this book gives you actionable steps you can take to do DI, today and right now, with or without technology. This book also focuses on DI processes: our emphasis here is on the sequence of steps to take within your organization to make better decisions with better outcomes.

Contents of This Book

In Chapter 1, we introduce you to DI. We present a brief history of DI and explain its benefits from several points of view. You can skip Chapter 1 if you want to get started with the DI processes quickly.

The remaining chapters are organized around nine DI processes, summarized in Figure P-3.

Figure P-3. The nine DI processes.

Chapter 2 gets you going with decision making. It covers the processes of creating an initial Decision Objective Statement and framing the decision design, including identifying the actions available and the desired outcomes.

Chapter 3 covers Decision Design, where you create your initial CDD.

In Chapter 4, you’ll investigate the technical and data assets that can support your decision. This is called Decision Asset Investigation.

Now it’s time to pull everything together to make the best decision you can. This is Decision Simulation, covered in Chapter 5.

Before you take action based on your decision, you’ll want to evaluate risks, sensitivities, and uncertainties. This is Decision Assessment, the topic of Chapter 6.

Now it’s time to take the action(s) you chose. In this book, we’re not going to tell you how to do things once you’ve made a choice—we figure you’re already pretty good at that. But in Chapter 7, we’ll describe Decision Monitoring: how you can use your DI assets to monitor the results of your action(s) (KPIs, intermediates, outcomes, and more) as they play out in reality so that you can make quick adjustments if things drift off course.

Finally, Chapter 8 covers what you do after the decision model has been used. In Decision Artifacts Retention, you ensure that as much of the decision-making effort as possible can be reused. In the Decision Retrospective process, you’ll assess and improve your decision-making processes for next time.

Conventions Used in This Book

The following typographical conventions are used in this book:

Italic

Indicates new terms, URLs, email addresses, filenames, and file extensions.

Constant width

Used for program listings, as well as within paragraphs to refer to program elements such as variable or function names, databases, data types, environment variables, statements, and keywords.

Tip

This element signifies a tip or suggestion.

Note

This element signifies a general note.

Using Supplemental Materials

Supplemental material (worksheets, etc.) is available for download at https://oreil.ly/DIH-supplemental.

If you have a technical question or a problem using the code examples, please send an email to [e-mail].

This book is here to help you get your job done. In general, if example code is offered with this book, you may use it in your programs and documentation. You do not need to contact us for permission unless you’re reproducing a significant portion of the code. For example, writing a program that uses several chunks of code from this book does not require permission. Selling or distributing examples from O’Reilly books does require permission. Answering a question by citing this book and quoting example code does not require permission. Incorporating a significant amount of example code from this book into your product’s documentation does require permission.

We appreciate, but generally do not require, attribution. An attribution usually includes the title, author, publisher, and ISBN. For example: “The Decision Intelligence Handbook by L. Y. Pratt and N. E. Malcolm (O’Reilly). Copyright 2023 Quantellia, L.L.C., 978-1-098-13965-0.”

If you feel your use of code examples falls outside fair use or the permission given above, feel free to contact us at [e-mail].

O’Reilly Online Learning

Note

For more than 40 years, O’Reilly Media has provided technology and business training, knowledge, and insight to help companies succeed.

Our unique network of experts and innovators share their knowledge and expertise through books, articles, and our online learning platform. O’Reilly’s online learning platform gives you on-demand access to live training courses, in-depth learning paths, interactive coding environments, and a vast collection of text and video from O’Reilly and 200+ other publishers. For more information, visit https://oreilly.com.

How to Contact Us

Please address comments and questions concerning this book to the publisher:

• O’Reilly Media, Inc.
• 1005 Gravenstein Highway North
• Sebastopol, CA 95472
• 800-889-8969 (in the United States or Canada)
• 707-829-7019 (international or local)
• 707-829-0104 (fax)
• [e-mail]
• https://www.oreilly.com/about/contact.html

We have a web page for this book, where we list errata, examples, and any additional information. You can access this page at https://oreil.ly/DI-Handbook.

For news and information about our books and courses, visit https://oreilly.com.

Find us on LinkedIn: https://linkedin.com/company/oreilly-media

Follow us on Twitter: https://twitter.com/oreillymedia

Watch us on YouTube: https://youtube.com/oreillymedia

Acknowledgments

This book would not have been possible without, first of all, the amazing team of employees and investors at Quantellia, who have stuck with us through the challenging task of creating a new discipline at the same time as we offer solutions into it. It’s been a long haul, but an epic experience. There are really no words to express our gratitude for our long-suffering supporters. Of particular mention is Mark Zangari, who, as the co-inventor of DI, is the “hidden gem” behind much of what we do: his insights are woven deeply into what you will read here.

We’ve also been blown away by the knowledge and experience of the team at O’Reilly, especially Michelle Smith, and Sarah Grey, who took the time to learn DI so deeply while editing our draft that we consider her a contributor to the discipline!

We’ve also been blessed with great colleagues with whom we have built and delivered DI solutions, including Jessica Jaret, Elizabeth Nitz, Dr. John Wooten, Skye Wiersma, and Katie Harder, and in earlier days Rick Ladd, Sammy Thomas, Jennifer Fruehauf, Margaret Johnson, Allie Golon, and Janet Nemmers. We are also particularly grateful to Erik Balodis, who contributed great ideas to an early draft of this book, resulting in this article.

Bolstering and supporting our work from outside of our organization have been friends and fans too numerous to mention. Worthy of special note are Jim Casart, Dr. David Roberts (a worldwide leader in bringing DI into academia, who also reviewed a draft of this book), Allan Frank, Robert Walker, Joseph Wiggins, VR Ferose, Håkan Edvinsson, James Parr and his team at Trillium (who are bringing DI into one of the most important domains of our time: climate resilience), Dr. Grady Booch, Dr. Cassie Kozyrkov, Jeffrey Williams, Tim McElgunn, and Linda Kemp. We got terrific feedback from a number of technical reviewers, and are grateful to Joshua Dejong, Dr. Roger Moser, Donald Farmer, Tobias Zwingmann, Anand Thakar, Jen Stirrup, and Jazmine Cable. The DI vendor community has also been of great support. Data Innovation.AI, Pyramid Analytics, Astral Insights, Agilisys, CModel, and C-Plan.IT are worthy of special note. We are also very grateful to our family members: Michael Malcolm, Dr. Annis Pratt, Dr. Faith Hopp, John Smith, Casper Smith, and Aspen Smith (and coauthor Pratt thanks her dog, Bowie, too), for their sacrifices that came from supporting this important endeavor. Co-author Pratt also thanks Landlocked Ales (whose Decision Fatigue beer flights fueled her writing).

¹ L. Y. Pratt, Link: How Decision Intelligence Connects Data, Actions, and Outcomes for a Better World (Bingley, UK: Emerald Publishing, 2019).