DecisionOS — Build · module
Readiness
Given what we have decided to do, the one thing that stops us — and what becomes the next constraint after we fix it.
When you reach for it
Immediately after a Decide-line run has produced an agenda, or at the start of a planning cycle when a CIO has to commit to a date in front of a board. Also, defensively, when somebody senior has asked whether you are ready and the honest answer is that nobody has checked.
What it asks you
9 blocks, in this order. Each one exists for a stated reason, and the run tells you that reason as you answer it.
- 01
The agenda
One to three things you are actually about to do. Readiness in the abstract does not exist — an organisation is not 62% ready, it is ready or not to do a particular thing at a particular consequence level — so everything below is measured against this list and against nothing else.
- 02
What this agenda demands
Three facts that set the bar. The requirement for every concern below is derived from these and printed as a rule, rather than being a maximum everyone is measured against.
- 03
Scope
Two questions asked once for the agenda. Both change who finds out when something is wrong, rather than how wrong it is, and both raise what the five concerns have to be.
- 04
Gate: Data
Whether the specific data these initiatives need can be reached, by a named route, at the cadence the decision requires. Three questions of fact and one date. None of them is a rating, and the concern is placed at the weakest of the three rather than at their average.
- 05
Gate: Operating model
Whether the workflow has a named owner who can change it, and whether decision rights for an agent acting inside it have been written down anywhere. Three questions of fact and one date. None of them is a rating, and the concern is placed at the weakest of the three rather than at their average.
- 06
Gate: Talent
Whether the roles the agenda implies exist today, and whether anyone could tell that an output was wrong. Three questions of fact and one date. None of them is a rating, and the concern is placed at the weakest of the three rather than at their average.
- 07
Gate: Architecture
Whether a system of this shape can be deployed, integrated and logged on the platform that exists, without a decision nobody has made. Three questions of fact and one date. None of them is a rating, and the concern is placed at the weakest of the three rather than at their average.
- 08
Gate: Control
Whether coverage at the level these consequence ratings require is in place, carried in from Governance where a run exists. Three questions of fact and one date. None of them is a rating, and the concern is placed at the weakest of the three rather than at their average.
- 09
Who answered
A CIO answering the data gate alone is a different artefact from the data owner answering it, and the report says which it is. Recorded, not judged — one person assembling all five is a legitimate way to run this and a normal one.
The method
How the result is produced
Each of the five concerns is placed by three recognition questions — can you produce it, who owns it, has it happened — and the concern sits at the weakest of the three, never at their average. A fourth question asks when the thing last happened, and caps the level: Operating means it runs on a cadence with a date for the last time it ran, so a gate claiming Operating with an older date is reported as Defined and the date is named. A concern whose three questions were not all answered is not placed at all, and while anything is unplaced the binding constraint is stated as provisional. The requirement for each concern is then derived from the agenda — the highest consequence level on it, raised a level if anything is customer-facing and a level again if deployment is organisation-wide, and floored where a concern is something nothing runs without. The binding constraint is the largest deficit against that requirement, never the lowest level. The second constraint is recomputed with the first resolved, and any concern that takes longer to fix than the binding one is flagged to start immediately.
What it connects to
Answers established elsewhere are carried in with their age shown, so a stale answer reads as stale rather than being silently reused.
The rest of the Build line
Operating Model
Who does what in this workflow, who approves what, what it costs to supervise, and who is accountable when the agent is wrong.
Data Foundation
Whether the data this initiative needs can be reached, is right enough, is permitted, and is fresh enough for the decision it feeds.
Talent
Whether we can staff and supervise what we have decided to do — and whether we still will be able to when the people who can catch the machine's mistakes have moved on.
Architecture
What has already been decided about how AI runs here, which of those decisions nobody actually made, and which will be expensive to reverse.