// The difference
These fields are related but not the same. Decision management, decision intelligence and decision science each do important work on the decision itself. Decision Engineering™ takes a different unit of analysis: the complete institutional decision, and the joins where intent is lost between them.
// The established fields
Any honest account of this space has to start with the people who built it. Three bodies of work matter most.
James Taylor, who leads Decision Management Solutions, is one of the original submitters of the Decision Model and Notation (DMN) standard and a leading authority on enterprise decision management and digital decisioning — the practice of modelling, automating and governing operational decisions at scale. With analyst Neil Raden of Hired Brains Research, he co-authored Smart (Enough) Systems (2007), one of the first books to make the case for automating the many hidden, high-volume decisions inside an organisation. This field is about making a decision reliable, transparent and repeatable once you know what it should be.
Decision intelligence connects data, decisions and outcomes — bringing analytics, decision modelling and feedback together so that choices improve over time. Its contribution is to treat the decision, not just the data, as the thing worth engineering for value.
Decision science studies how people should and do make choices. The cognitive psychologist Gary Klein reshaped this field by pioneering Naturalistic Decision Making and the Recognition-Primed Decision model — showing, in Sources of Power, how experts such as firefighters and commanders actually decide under pressure. This work is about the human act of judgment.
// Where Decision Engineering™ differs
Each of those fields is strong within its own frame, and each takes the decision — or the decider — as its unit of analysis. Decision Engineering™ starts one level up, at the institution, where a material decision is rarely made by one person or one model. It is made across a chain: a board sets purpose, policy translates it, data and technology encode it, and people or automated systems execute it, sometimes thousands of times a day.
At every transition something is added and something is lost. The question Decision Engineering™ asks is not "is this the best decision?" but "is the decision the institution is executing still the decision it authorised — and can it prove it?" The risk it addresses lives in the joins between separately governed areas, where intent, authority and accountability quietly separate. That is why a decision can be operationally correct, technically compliant, and still completely different from what leadership intended.
It is developed through the Decision Integrity Chain™ (an eight-layer map of a decision's journey), The Fiduciary Gap™ (the distance between who decides and who is accountable), and replayability (whether a decision can be reconstructed while there is still time to change what follows).
// At a glance
| Discipline | Primary question | Associated with |
|---|---|---|
| Decision management | How do we model, automate and govern a decision reliably at scale? | James Taylor (DMN, Decision Management Solutions); Neil Raden (decision automation, Smart (Enough) Systems) |
| Decision intelligence | How do we connect data, decisions and outcomes to improve choices? | An emerging field across analytics and decision modelling |
| Decision science / NDM | How do people — especially experts — actually make decisions? | Gary Klein (Naturalistic Decision Making, Recognition-Primed Decision, Sources of Power) |
| Decision Engineering™ | Does institutional intent survive from purpose to execution — and can it be reconstructed? | Deepak Aggarwal (Decision Integrity Chain™, The Fiduciary Gap™) |
These are complementary, not competing. Decision management and decision intelligence make the decision better; decision science explains the human judgment inside it; Decision Engineering™ verifies that the institution is executing the decision it authorised.
// A worked distinction
Consider a failure where a decision crossed boundaries that no single discipline held as a complete object. On 1 August 2012, Knight Capital deployed new software to seven of its eight order-routing servers; the eighth still ran code from 2003, where a reused flag meant something entirely different. Every component behaved according to the code actually running on it — but not according to the decision anyone believed had been authorised. In forty-five minutes the firm lost about US$440 million.
Decision management could model and automate the trading rules. Decision science could explain the judgments of the people involved. Neither held the decision as a single institutional object that could be traced from intent to execution across the join where it broke. That gap — between what was authorised and what was executed, and whether the institution could reconstruct it in time — is the object Decision Engineering™ was built to examine.