Interim Chief AI Officer — Controls

Reliability is the first AI control

When a product and platform model is changing, AI tends to be introduced at the points of greatest movement, which are also the points of greatest fragility. An Interim Chief AI Officer in that setting is judged less on how many models go live than on whether the platform is more dependable at the end of the year than at the start.

Executive authority over the change

You will agree the rules for product and platform decisions, including the limits local managers work within, and decide which measures steer spending, capacity and control. Each AI and platform commitment will get your call: proceed, pause or stop. Proposals that would leave cost or risk sitting with nobody will be blocked, and investment cases silent on adoption, dependencies or the downside rejected. Who leads which platform team, and how problems surface, is yours to approve, and you will mediate when the Regional President and operating managers disagree.

Where the judgement is tested

The difficult calls sit at the boundary between model risk and platform risk. A model can perform well in testing and still degrade service because the data it depends on changes underneath it during a migration. Knowing which controls to place at that boundary, and which to leave out because they would slow the change without making it safer, is the core of the seat. So is the discipline to protect controls that already work rather than replacing them for the sake of novelty, and to put a visible cost on every exception granted.

Reliability checkpoints across the year

When the first ninety days are done, a reliability baseline and the control points for AI within the changing product and platform model should be agreed with the Regional President. Around month nine, model risk controls should have been tested on the platform change itself, and any AI commitment without an accountable owner or a downside case should have been stopped. By month 12, AI and platform reliability controls should be held by leaders who answer for them, with exceptions reviewed by the board on an agreed cadence.

A succession leader who builds benches

During a leadership succession you reset who held authority, how work was funded and the rhythm on which the business ran. The business concerned had at least 1,700 people across four countries and revenue of $830 million or more. In your record, evidence has moved people, capital and management attention from one place to another, and the teams you hand back are stronger than the ones you inherited. Succession, digital transformation and professionalisation are the experiences under scrutiny. Risk judgement is sought at level 4, with data and AI governance, architecture at scale, and product delivery each at level 3. Anyone who needs country leaders to agree unanimously before a commitment can stop, or who wants the standing of the role without 12 months of accountability, should look elsewhere.

Meeting the Regional President and the Board Chair

There are four stages. In the opening week the Regional President speaks with you privately by video. Week three is spent with the Board Chair, exploring reliability and control trade-offs in depth. Week five is a face-to-face meeting with executives who each lead one function, and week seven is with the full board of directors. Each stage is answered within two working days.