An AI position by the quarter
Illustrative example built on simulated data. Not a client engagement.
The situation
The board of a simulated logistics company asks management for an AI position within one quarter. The first draft is a policy document: principles, a risk statement, an ambition. The board sends it back. It wants to know what the company will fund, what it will not, and why.
The evidence
Eight process areas were assessed in two working sessions, covering 2,900 monthly effort hours. 76% of the evidence needed was available. Each process area was scored for AI suitability, oversight requirement and value, with every assumption written down.
What the assessment showed
Two process areas had a clear case, one had a case only under close oversight, three were not suited to AI at their current maturity, and two lacked the evidence to decide either way.
The decision
- Invest now freight invoice audit, delivery exception handling
- Governed pilot customer booking enquiries
- Defer route planning and warehouse slotting, until their data is complete
- Keep human-led driver safety reviews, customs declarations, key account negotiations
Estimated value band: €350–700k a year, assumptions stated.
The board receives the Executive Operating Report and the decision record. Reassessment in six months.
What would change it
The position is wrong if freight volumes fall by more than 20%, if the booking pilot's error rate exceeds 2%, or if a process is classified as high-risk under applicable AI regulation. Each of these is a named trigger in the record, so the board knows in advance what would reopen the decision.