Boards, trustees & committees
This lens applies the course to boards, trustees, and committees. Complete the eight core modules, and use the sector-specific examples and prompts here to ground them in board governance. It does not repeat the course — it tailors the lens and points back to the core material, where the substantive teaching, simulations, and self-checks live.
Work the core modules with a board hat on
Where the risk shows up for boards
Board governance concentrates the course's five risk categories into a handful of recurring, high-consequence records. The same artefacts that make a board efficient — summarised packs, delegated approvals, terse minutes — are the ones most likely to fail under later challenge if their provenance is not verifiable.
Mapped to the five risk categories
- Integrity — minutes contested after the fact; whether the approved version of a resolution can be shown to be the unaltered one. Was this what the board agreed, or what was later written down?
- Jurisdiction — restructuring deliberations and conflict-of-interest records sitting in vendor-controlled tools, on infrastructure outside the board's effective control, under terms that can change unilaterally.
- AI reuse — board-pack AI summaries that omit dissent, flatten nuance, or soften a risk; the question of what the summarisation model was permitted to do with confidential papers.
- Contestability — voting trails and dissent pathways: can a director's recorded dissent, or the chain of a delegated-authority decision, be reconstructed and stood behind years later?
- Trust — whether the board relies on trust in people and vendors, or trust in the record itself, when a decision is questioned by a regulator, member, or court.
Worked examples
Three board-shaped scenarios that the core modules equip you to handle. Each maps a concrete failure mode onto the verifiable-record discipline taught in the course.
What to emphasise from the core modules
The full teaching is in the modules; for boards, three pieces carry the most weight.
Board emphasis
- Module 6 — the human-in-the-loop boundary matrix. Apply it directly to the board secretariat AI: which summarising, drafting, and routing tasks the model may perform unaided, which require a named human in the loop, and which it must never touch. This is the control that turns "the AI helped" into a defensible boundary record.
- Module 7 — deliberation-chain reconstruction. Use it on a contested resolution: rebuild the papers, options, conflicts, amendments, votes, and dissent into an attributable, verifiable chain. This is what answers the "challenged two years later" scenario above.
- The capstone — board action statement. Convert the lens into commitments: which board records become sovereign, who owns the boundary matrix for the secretariat AI, and how directors record dissent with a verifiable trail.
Board discussion prompts
- If your last contested decision were challenged today, could the board prove how it was formed — papers, conflicts declared, options weighed, and the vote — rather than only what was minuted?
- What is your board secretariat AI permitted to do with confidential board papers, and who would know if a summary had omitted a dissenting view?
- How does a director currently record formal dissent, and could that record be quietly altered or dropped before it reaches the approved minutes?
- For delegated-authority decisions taken between meetings, where does the verifiable trail of who decided what, and under which delegation, actually live?