Strategy questions for the board agenda
- Which mission outcomes could AI improve in 12–24 months?
- What will we refuse to automate?
- What budget and skills are we willing to fund?
- How will we know if shadow AI is spreading?
- What metrics will management bring quarterly?
These questions keep the conversation on mission and risk—not on brand names from the latest conference booth.
Risk appetite in plain language
State appetite clearly. Example: “We accept AI drafting for public communications with human edit; we do not accept AI in safeguarding decisions.” Clarity frees staff to move inside bounds. Silence forces them to invent bounds—or ignore them.
Magnifica Humanitas’s insistence that technology is never neutral belongs in the risk discussion: choosing defaults is still a choice.
Oversight cadence
Quarterly: inventory changes, incidents, pilot results, vendor issues. Annually: principles and budget. Avoid both neglect and micromanagement. One inventory and one decisions log beat a thick binder of unread policies.
Mission test
Every major AI initiative should answer: how does this serve persons we are responsible for, without instrumentalising them? If the answer is only cost-cutting, re-examine design and safeguards. Efficiency that hollows dignity is not stewardship.
What not to ask the board to do
Do not ask trustees to choose model brands by logo. Do not bury them in token-price spreadsheets without a mission narrative. Do not present AI as inevitable destiny. Present options with costs, risks, refuse-lists and a recommendation. Boards govern; management executes.
Bring one pilot story—success or failure—each quarter. Stories teach faster than dashboards alone.