The scoring and sequencing instrument used in AI Opportunity Audits. It takes the backlog of AI ideas every organisation now has, scores each on value, effort and readiness against shared criteria, and returns a build order leadership can fund.
The usual picture: two dozen AI ideas from every department, no shared way to compare them, and a choice made by whoever sponsors loudest.
The result is the pilot graveyard: three initiatives started at once, none resourced to finish. A year later the organisation decides AI "didn't work here", when what actually failed was the selection, which was never made on purpose.
Leadership commits to the scoring model first, so the ranking that comes out is theirs. Value is scored on the real mechanism (hours returned, revenue protected, risk retired), effort includes the change management AI programs forget to price, and readiness checks whether the data exists and the process is stable enough to encode.
Scores never stand alone. Written overrides do the judgment work: a high scorer waits while its team reorganises, a modest one moves up when it feeds a bigger build later, and anything that fails the readiness gate leaves the list.
The deliverable is a workbook, not a slide. The model, the scores and the reasoning stay with the client.
Four stages between the idea list every organisation has and the build order almost none of them do. Click a stage.
Move the sliders. Ideas that fail the readiness gate drop out, whatever their value.
The cheapest kill an AI program will ever make is at the selection stage.
Every organisation has the list. Few have made an explicit selection.
Strong value on weak data is the most common way initiatives die.
Agreed before scoring, the model is leadership's logic, not the consultant's opinion.
The client re-runs it as priorities change.
If the honest answer is "long" and "nobody", this instrument was built for you. Start with the three ideas your teams argue about most.
📅 Book a call