B
Business
Why are we doing this? What value or outcome are we expecting? What has to be true for the decision to make sense?
Our philosophy
Trust,
but verify.
Good governance does not start from distrust.
Boards have to rely on management, advisers, vendors, auditors, and specialists. No board can independently recreate all of the work beneath every decision.
The question is not whether to trust. It is where independent verification matters.
Oversight
Reliance is necessary. Oversight is different.
A board can receive good information, ask sensible questions, and make a sound decision. But implementation happens after the meeting.
Systems change. Vendors change. Organisational use expands. Assumptions that supported the original decision may also change.
None of this automatically means something has gone wrong. It does mean that, at important points, the board needs a way to check whether the reality still matches the decision.
Verification is not a substitute for trust. It is part of responsible oversight.
The board does not need to know everything.
It needs to know what matters.
Business, assurance, technology
Three perspectives. One decision.
AI decisions rarely belong to one discipline. Looking at them through only a technology, risk, or commercial lens leaves part of the picture out.
A
Assurance
How would we know whether it is working as intended? What could materially undermine the decision? What evidence should the board be able to see?
T
Technology
What is the AI actually doing? What data, systems, vendors, and dependencies sit underneath it? What can change after approval?
Bringing the perspectives together
Board questions before technical answers.
Business, assurance, and technology matter because they help directors ask better questions about the decision in front of them.
Business
What are we trying to achieve, and what would make this worthwhile?
Assurance
How will we know whether what we approved is what actually happened?
Technology
What sits underneath the decision, and what could materially change?
Good AI governance should make important decisions clearer, not bury them under more governance.
Frameworks, policies, controls, and technical detail have their place. The board's job is different: understand the decision, challenge what matters, and be able to see whether implementation remains aligned.
Trust the people doing the work. Verify what matters.
That is the thinking behind the AI Oversight Stocktake.
The Stocktake →