From the Boardroom to the Factory Floor.

01Two rooms, one week

The first room runs on judgement. The inputs are incomplete, the stakes are strategic, and the useful contribution of a system is not to decide anything — it is to make the basis for a decision visible: what is known, what is assumed, what changed since last quarter and which of those numbers is softer than it looks.

The second room runs on repetition. The input is a physical object or a signal, the stakes are per-unit and cumulative, and the useful contribution is a decision made consistently, at line speed, with a known error profile and a clear rule about what happens when the system is not sure.

02What changes between them

Almost everything at the surface. Latency budgets that differ by six orders of magnitude. Interfaces designed for one considered decision a quarter versus none at all. Failure that means a poorly-informed strategy versus failure that means scrap, or a line stopped, or somebody in the wrong place at the wrong moment.

Between the two sit the enterprise workflows that most organisations actually run on — the approvals, the case handling, the knowledge nobody can find, the reconciliations that consume a team's month. Less dramatic than either extreme, and usually where the largest quantity of recoverable time is buried.

• 01Boardroom: incomplete inputs, high consequence, one decision at a time.
• 02Enterprise: high volume, high friction, decisions nobody enjoys making.
• 03Operations: real time, physical consequence, no tolerance for drift.

03What does not change

The method is identical, which is the part people find surprising. Go and see the work. Find where the time and the risk actually sit rather than where the org chart says they should. Establish what 'working' means in numbers before building anything. Design the unsure case first. Instrument it so that its behaviour can be inspected later by somebody who was not there.

So does the engineering underneath. A vision model on a line and a decision system in a boardroom both need versioned data, controlled deployment, monitoring that notices drift, and an owner after go-live. The consequence differs by orders of magnitude. The discipline does not differ at all.

The distance between an executive dashboard and an inspection camera is enormous in every respect except how you have to build them.

04Why we keep the whole span deliberately

It would be simpler to specialise at one end. We keep both because each keeps the other honest. Operational work is a permanent corrective against systems that are impressive in a demonstration and unusable at three in the morning. Strategic work is a permanent corrective against building something technically excellent that nobody needed.

It is also the only way to answer the question organisations actually have, which is rarely 'can you build this model' and almost always 'what should we do about this, and what will it take'. Answering that requires having stood in both rooms.