DecisionOS — Decide · module

Use Cases

Which AI ideas do we do, in what order, and which are we explicitly not doing.

When you reach for it

Two weeks after an ideation workshop produced thirty ideas and no decision, or the week before a steering committee where you must arrive with a sequence rather than a menu.

What it asks you

4 blocks, in this order. Each one exists for a stated reason, and the run tells you that reason as you answer it.

01
The candidate set
Name what is actually on the table. Between three and fifteen — fewer than three and there is nothing to compare, more than fifteen and this is portfolio work rather than a decision.
02
Five facts per candidate
Facts, not ratings. Each one is something a person can actually know, and 'don't know' is a real answer that changes the output rather than scoring as a middle value.
03
The forced trade-off
Three candidates at a time: which would you fund first, and which last. Every comparison forces a last choice, so rating everything highly is structurally impossible.
04
The kill test
Assume eighteen months have passed and this was judged a failure. Asking why it failed surfaces risks that 'what could go wrong' does not, because it licenses pessimism.

The method

How the result is produced

Two orderings are computed from different sources. The evidence order comes from a published rule set: a value signal built from frequency, process ownership and demand pull, and an executability state built from the data situation and downgraded where ownership is unclear. The preference order comes from best-worst scaling over the trade-off comparisons you completed. Neither is a score out of 100, and there is no model in between. Each candidate is then placed in exactly one band by an ordered rule set, first match wins, and the rule that placed it is shown against it. Where the two orderings disagree by three places or more, the gap is reported with the facts behind it.

What this is not

This is not an assessment of your organisation. It ranks the candidates you entered, using the facts you gave about them, and nothing else.

This is not a business case. A candidate in 'Do now' has a clear path, not a proven return — that is the ROI module's job.

This is not a discovery exercise. Candidates nobody entered do not appear, and the run cannot tell that they are missing.

Where it stops

Everything here is what you told us. Nothing was read, inspected, profiled or connected to.

There is no peer comparison behind any statement in this report, and none is implied. We will not show one until there is enough real data to state an n.

Frequency and consequence are your estimates. The ranking is sensitive to both, and the sensitivity section names which candidates would move.

This run reflects the answers of whoever completed it. A second respondent answering the same questions frequently disagrees, and that disagreement is usually the finding.

Where no Governance run exists, every C4 and C5 candidate is treated as having no coverage evidence — because from here, it does not.

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.

Reads saved results from

GovernanceOpen

Its results carry forward into

The rest of the Decide line

Portfolio

Which of the AI work already running should be scaled, sustained, merged, fixed or stopped — and what stopping frees up.

Open

ROI

What this initiative is actually worth, as a range, and which single assumption the answer depends on.

Open

Vendors

Which of these options to choose — including building it ourselves and doing nothing — and what the choice actually rests on.

Open

See what it costs All twelve modules