Board
The board sits outside the day-to-day model. It has one line into the model, through Portfolio governance and AI risk, and no direct role in the daily loop between functions and customers.
What the board holds
- A named AI owner. One person at board level is accountable for AI and for the operating model. This might be a Chief AI Officer or an existing board member such as the CEO, MD or COO.
- AI risk on the risk register.
- Approval of the AI governance policy.
- Final accountability for the Portfolio.
Why AI cannot belong to one function
AI runs through every part of the model: delivery, pricing, finance, contracts, compliance and infrastructure. If one function owns it, the business sees AI through that function's priorities. Ownership therefore sits at board level, and day-to-day responsibilities are spread across the functions through a RACI matrix, with each function accountable for AI within its own area.
The final escalation point
The board-level AI owner resolves the tensions that the functions cannot settle between themselves. The main routes are:
| Disagreement | Who decides |
|---|---|
| A vendor will not accept terms that protect the business | Board-level AI owner, escalated by legal |
| Legal and compliance disagree on whether a risk is acceptable | Board-level AI owner |
| IT and a requesting team cannot agree on a security, compliance or regulatory issue | Board-level AI owner |
| IT and a requesting team cannot agree on cost | Finance, then the board-level AI owner if finance will not accept the risk |
| Portfolio trade-offs between functions | Portfolio executive sponsor, with the board accountable |
| Other parts of the business bypass the skill matrix | Centre of Excellence Director |
Evidence status
Evaluated against an illustrative scenario. Not yet tested in a live organisation.