Board AI Governance: What Directors Need to Ask in 2026
Boards do not need to manage AI projects, but they do need to understand where AI creates strategic value and material risk. These are the questions directors should be asking.
By Sophie Keller, Women in AI Editorial Fellow ยท 31 August 2026
Artificial intelligence is now important enough to appear regularly on board agendas, but that does not mean directors should become project managers for AI.
The board's role is oversight. It should understand how AI affects strategy, risk, capital allocation and organisational capability, then challenge management on whether those issues are being handled coherently.
Where is AI creating strategic value?
Boards should ask which AI investments are expected to change the economics or competitiveness of the business.
A long list of pilots is not a strategy. Directors need to understand which use cases matter, what outcomes they are supposed to improve and how management will know whether the investment is working.
Who is accountable?
Responsibility for AI often becomes fragmented across technology, data, legal, risk and individual business units.
The board should know who owns enterprise AI strategy, who owns material AI risk and how disagreements are escalated.
Frameworks such as NIST AI RMF and ISO/IEC 42001 both emphasise governance and organisational accountability because technical controls alone are not enough.
What systems could materially harm the company or its stakeholders?
Boards do not need to review every low-risk internal tool.
They should understand the systems that could create material financial, regulatory, reputational or customer impact. That may include high-risk automated decisions, sensitive data use, customer-facing agents or systems with significant autonomy.
How is management measuring AI performance?
Directors should challenge vanity metrics.
Number of licences, prompts or pilots can indicate activity. They do not prove value.
Ask for evidence tied to business outcomes: cycle time, quality, revenue, cost, customer experience or risk reduction.
Is the workforce ready?
Microsoft's 2026 Work Trend Index argues that organisational conditions have a major influence on AI impact. That makes workforce capability a board-level issue when AI is central to strategy.
Directors should ask whether managers understand how work is changing, whether employees receive useful training and whether the company is building the skills it will need rather than only buying tools.
Does the governance model evolve as AI becomes more autonomous?
AI agents create new questions about permissions, monitoring and human approval.
A board should understand whether management has considered what systems are allowed to act without human intervention and where stronger controls apply.
These board-level questions will be part of the leadership and governance conversations at the Women in AI Global Summit in London, where executives, technical leaders and policymakers will be able to compare how different organisations are approaching AI accountability.
The board does not need to know everything
Directors do not need to understand every model architecture. They do need enough AI literacy to recognise weak answers.
Good oversight comes from asking whether the company knows what it is trying to achieve, what could go wrong, who is accountable and what evidence management uses to make decisions.
That is not a new governance discipline. AI simply makes the questions more urgent.