The Decision Intelligence Handbook (L. Y. PrattN. E. Malcolm)（Z-Library）

# The Decision Intelligence Handbook ## Practical Steps for Evidence-Based Decisions in a Complex World ### 【One-Line Pitch】 A practical, roll-up-your-sleeves guide to Decision Intelligence (DI)—a methodology for making better, evidence-based decisions by connecting actions to outcomes through causal decision diagrams. Essential reading for executives, data/AI consultants, and anyone who wants to move beyond "data-driven" buzzwords to actually design and improve decisions. --- ### 【Book Arc】 - **Opening (~0%–10%)**: Introduces DI as a discipline—what it is, why it matters, and how it differs from plain data analysis. Sets up the core concept of the causal decision diagram (CDD) and the nine DI processes that structure the rest of the book. - **Early (~10%–23%)**: Makes the case for DI through real-world examples and common organizational pain points—from climate data going unused to KPI tracking that doesn't tell you what to do when things go wrong. Establishes the "human-in-the-loop" philosophy and the DI Maturity Model. - **Early–Middle (~23%–42%)**: Begins the hands-on work with Phase A: Decision Requirements. Covers the Decision Objective Statement (the trigger that starts a DI initiative), convening the decision team, and the critical balance between information, authority, and responsibility. - **Middle (~42%–48%)**: Deepens the decision framing process—how to structure the decision, handle multi-level organizational decisions, and establish a decision glossary. Emphasizes that decision models support intuition rather than replace it. - **Late (~48%–end)**: Walks through the remaining DI processes—designing the decision, connecting data and AI, and building continuously improvable decision assets. The excerpts suggest the book culminates in practical implementation guidance for organizations and consultants. --- ### 【Key Takeaways】 - **DI is about action-to-outcome decisions** (Early): The core question is "what actions can I take today to achieve my desired outcomes tomorrow?" This distinguishes DI from generic data analysis or KPI tracking. - **The causal decision diagram (CDD) is the central tool** (Early): A drawing that maps how potential actions connect to desired outcomes, making the decision logic explicit and testable. You can build your first one in about 20 minutes. - **DI is a human-in-the-loop methodology** (Early): It applies to decisions made by people (with technology support), not fully automated systems. It's an intelligence augmentation approach—working *with* AI, not handing decisions over to it. - **The Decision Objective Statement anchors everything** (Middle): A simple trigger—often an email starting with "please help me to decide…"—that frames the decision and keeps the team focused on the original request from the decision customer. - **Balance information, authority, and responsibility** (Middle): Decision models make data and knowledge explicit, but the decision customer must weigh model recommendations against intuition—especially when they conflict. Models improve but don't replace judgment. - **DI is a systemic fix, not a silver bullet** (Early): It builds on over a century of management innovation rather than replacing it. The nine processes are learnable one step at a time. - **LLMs are a "super Google" for DI** (Early): Large language models can surface actions, outcomes, and unintended consequences that people hadn't considered—reducing tunnel vision—but they don't do action-to-outcome simulations, so they complement rather than replace DI methods. - **DI delivers value across decision types** (Early): From one-of-a-kind strategic decisions to repeated tactical decisions in business processes, and even in novel situations where organizations are "working in the dark." --- ### 【Reading Tips】 - **Skim Chapter 1 if you're in a hurry**: The authors explicitly say you can skip the introductory context and go straight to Chapter 2 to start the hands-on work. Return to Chapter 1 later for the maturity model and history. - **Do the "Build Your First CDD" exercise**: The authors promise you can create your first causal decision diagram in about 20 minutes. Actually doing this will make everything else in the book click. - **Pay close attention to the Decision Objective Statement examples** (Middle): The book provides several concrete examples—from selling a company to sweet potato farming—that illustrate how to frame decisions properly. Study these before writing your own. - **Watch for the balance between model and intuition** (Middle): The discussion of information, authority, and responsibility is subtle but crucial. This is where DI differs from naive "trust the model" approaches. - **Note the LLM integration points**: The authors mention LLMs as collaborators throughout the DI processes. If you're an AI practitioner, look for where these integrations are described—this is a differentiator for the methodology. --- ### 【Coverage Limits】 The excerpts cover the book's introduction, the DI Maturity Model, and the beginning of Phase A (Decision Requirements). Detailed coverage of the remaining eight processes—decision design, data connection, and implementation—is not included in this guide's source material. --- ###