Most failed AI projects were not badly built. They were badly chosen.

The team picks the process that is most annoying, or most visible to leadership, rather than the one most likely to work — and discovers the mismatch nine months and a six-figure budget later.

Choosing well is a far cheaper problem to solve than building well. This seven-page worksheet ranks candidate processes against the six factors that actually predict whether an automation effort succeeds, so the conversation stops being about enthusiasm and starts being about evidence.

What is inside

A six-factor scored assessment — volume and repetition, rules over judgement, input quality, cost of being wrong, measurable outcome, and organisational readiness. Eighteen statements, scored out of 36.
Section subtotals, because automation efforts fail at their weakest factor rather than their average — your lowest score is the finding
Four verdict bands telling you whether to proceed, prepare, wait, or pick a different process entirely
A worked example scoring three real candidates, where the winner is the one nobody had nominated
Four objections you will hear and what each one actually means
The shape of a sound first ninety days, from validating inputs to deciding honestly at quarter end

Who this is for: operators and technology leaders with two or three candidate processes in mind who want to rank them on evidence before committing budget.

Get the worksheet

Prefer to skip the form? Download the worksheet directly (PDF).

General business guidance only. This worksheet is a planning aid, not professional, legal, financial or regulatory advice, and it does not assess compliance with any AI-specific regulation or standard. Automation decisions involving personal data, employment outcomes, credit, health, safety or other regulated activity carry obligations this scoring does not evaluate — take qualified advice on your own circumstances before proceeding.