Interim Chief Technology Officer — Scale and AI Governance

A technologist who will switch a model off

The person for this seat has put models into production and has also taken them out. That second experience matters more. In insurance, AI increasingly touches pricing, fraud detection, claims triage and customer contact, and each use raises questions technologists often leave to someone else: on what evidence does this model keep running, who may override it, and when does it retire?

You have led technology for organisations with revenue of $800 million or more and 1,300 or more people in at least two countries, and your experience of geographic expansion, digital transformation and professionalisation has taught you that models behave differently across markets, datasets and teams. The Group Chief Executive Officer expects risk judgement at level 4; architecture at scale, product delivery and the governance of data and AI are expected at level 3. You have carried a contested downside in front of a chief executive at full strength, and you have withheld money from cases that could not name their users or their risks, then rebuilt the technology capacity to do the work properly.

The CTO's powers over models and money

The CTO here is an executive, not an adviser. You will set how technology and AI decisions are made and the scope local technology leads keep for themselves. You will define the evidence that governs spending, engineering capacity and control across markets; for AI, that means the standard a model must meet to be deployed, to stay deployed, and to be overruled by a person. Every inherited commitment will continue, pause or retire on your decision. You will turn down proposals that leave cost or risk with a team that never took it on, and investment cases that ignore whether the business will use the tool, what it depends on and what it does when it is wrong. You will approve the leadership team's design and the route for raising problems, settle disputes with business heads before budgets are spent, and decide how often model performance, breaches and pending choices are reported upward.

Inventory, retirement, routine

Within ninety days there should be an inventory of AI uses, agreed evidence standards and override controls, and an accountable owner for every model in use. By around month nine, retirement criteria should have been applied to existing models, and AI commitments that cannot be supported should be paused or retired. At month 12, AI governance should run to a reporting cadence the board can test for itself, with owners in place who will hold it after the interim period.

Four stages, two working days

It starts with the Group Chief Executive Officer, who meets you by video in confidence during week one. The Board Chair takes the third week, working through a live model question with you. Week five gathers executives from each function to meet you face to face, and in week seven you meet the directors together. Each stage earns a written reply within two working days.