The seat · present in all three doors
Who answers to the board for what your AI decides?
AI moved faster than the org chart: it already decides and it still has no owner.
A named head of AI, under a written mandate: sets the portfolio, governs the risk and answers for what your systems decide.
Also searched for as part-time Chief AI Officer, interim head of AI or outsourced AI leadership.
The five decisions
The five decisions someone with a name has to be able to make about your organization's AI. An organization that has them assigned governs its AI with the same clarity as the rest of its operation:
- What gets in. Which use cases get approved, which get dropped, and by what written criterion.
- What gets switched off. Who can stop a system that decides badly without asking three departments for permission.
- With whom. Which vendors and which models are accepted. And what happens the day one of them changes without notice.
- With what evidence. What gets recorded for each automated decision so it can be reconstructed for a third party.
- To whom. Who all of the above is accounted for to, and how often.
The second is the one that brings the most peace of mind when it is settled in advance.
Before you read any further
Few mandates, on purpose.
Mandates are limited: the firm holds one or two seats at a time, counting technology leadership ones. If your organization can have a full-time head of AI, it should.
The firm holds two mandates at a time at most, counting technology leadership ones. That is not a courtesy: it is the shape of the product. If we say yes, it is because the seat is open.
Where this fits
Two seats of the same leadership.
Both are held under a written mandate and share the same cap. What changes is what the role answers for.
If what you are looking for is full technology leadership, see the Fractional CTO.
How the seat gets taken
- The mandate is agreed in writing: what the role decides, what it consults, and who it answers to.
- An inventory is taken of what already decides in your organization and who answers for each one today.
- The AI portfolio is set: what scales, what stops and what gets dropped, with written criteria.
- Risk is governed against a named framework, with the evidence a supervisor would ask for.
- The role reports to the board on a fixed cadence, and what must remain is installed before the mandate ends.
What stays in place when the mandate ends
- The written mandate, stating what the role decides and who it answers to.
- The inventory of what your AI decides, with one accountable person per system.
- The prioritized portfolio, with what scales and what gets switched off.
- Risk governance in place and accountability to the board.
This sounds like you if
- Systems are deciding in processes that matter and accountability is spread across departments.
- The committee asks who answers for AI and the answer depends on who replies.
- There is a technology leader, and that agenda is already full with everything else.
The proof that applies here
- 250,000 incidents analyzed before automating, and 700 agents in production under our own leadership.
- Eight years answering to a financial regulator without a single finding.
- Leadership of an operation of more than 150 people with its own budget.
What it is and what it is not
This is the function, not advisory: someone with a name who decides what scales and what gets switched off, and who answers to the board for what the systems decide. Whoever takes the seat answered to a supervisor for eight years for systems that had to work.
The person who listens, the person who leads and the person who signs are the same person.
If one of your systems decides wrongly about a customer tomorrow, who stops it and who answers to the board?
If the problem is a different one