For the next significant AI proposal, ask management whose experience could change the recommendation. Then ask how that experience reaches someone who can act on it.
A useful board paper can place relevant customer concerns, employee observations, and operating exceptions alongside performance measures. Our recommendation is to show what people are encountering, how the evidence was gathered, what remains unknown, and the response management proposes. Use this approach within the organization’s existing reporting and decision arrangements.
NIST’s AI Risk Management Framework Playbook provides a relevant starting point. GOVERN 5.1 addresses gathering and using feedback from people outside the team that developed or deployed the system. Its discussion connects engagement to the system’s purpose and risk and recommends appropriate expertise in participatory and qualitative methods. This is guidance for AI risk management, rather than a prescribed board-report format. Read GOVERN 5.1.
Define the experience you need to understand
Start with one consequential use of the technology. Ask which people could encounter its decisions or outputs and what management needs to learn from them.
Consider a hypothetical AI-assisted customer service process. Relevant questions might concern whether customers can obtain a correction, whether staff can identify incomplete answers, and what happens when a request falls outside the expected pattern. The questions should follow the actual use case.
Ask management to identify which existing channels could provide useful evidence. Complaints, service reviews, frontline observations, and targeted conversations are possible inputs. Have the paper explain why the selected channels fit the question rather than simply list every available source.
Show how a concern became a reported theme
For each material theme, request the period covered, the collection method, and an explanation of how the underlying observations were grouped. If a count is shown, ask what was counted and whether repeated reports of the same event have been separated from distinct events.
Keep the distinction between a reported experience and an established explanation visible. A customer may report receiving an incorrect answer. Investigation may still be needed to determine whether the cause involved the AI system, source information, a handoff, or another part of the service.
A board paper could summarize a concern as an issue for investigation, with management’s current assessment and the next evidence needed. It should avoid converting a plausible explanation into a settled finding.
Make missing perspectives visible
Ask who could be affected but is unlikely to appear in the evidence collected. Consider, for the particular use case, whether the channel is accessible to people with different language, access, or support needs. Ask whether employees have a suitable route to raise concerns about the work they are expected to perform.
NIST’s GOVERN 5.1 suggestions include considering feedback across intended users, including historically excluded populations and people with limited access to technology. The practical board question is which relevant perspectives management has sought and which remain missing. See NIST’s suggested actions.
Treat complaint themes as signals whose reach needs examination. Do not infer that a small set of comments represents everyone affected, or that an absence of complaints establishes an absence of problems. Ask what additional inquiry would be proportionate before drawing a wider conclusion.
Protect the people behind the evidence
Request aggregated or carefully deidentified reporting wherever it can answer the governance question. Avoid including names, account details, or unnecessary personal histories in a general board paper. Ask the responsible privacy or legal specialist to advise on what information may be collected, retained, and shared in the organization’s context.
Where a specific case requires closer review, use the appropriate restricted process and explain why that level of detail is necessary. Consider whether a small group or distinctive story could identify someone even after a name is removed.
These are recommended safeguards to examine with the responsible specialists. They do not establish that a particular reporting practice complies with applicable law.
Connect the evidence to a decision and an owner
For each significant issue, ask management to state its assessment, the action within its authority, and the condition that would bring the matter to the relevant committee or board. Name the person responsible for follow-through and identify the evidence expected at the next review.
In the hypothetical customer service example, feedback about unsuccessful correction requests might lead management to inspect the correction process before recommending wider deployment. Further inquiry might support a change, establish that another response is appropriate, or leave an important uncertainty unresolved. The board should be able to see which conclusion management reached and why.
Ask what changed because the evidence was considered. If management decided against a proposed response, request the rationale and any remaining exposure. Keep the inquiry open where the evidence is insufficient for closure.
Start with a bounded review
Choose one existing report and one material use case. Ask management to bring a concise account of the relevant experience, the limits of the evidence, and the proposed response. Use the discussion to determine whether the existing process needs an adjustment.
NIST describes its Playbook suggestions as voluntary and selectable, not a checklist to complete in full. The reporting approach here is an editorial application, with no claim that it guarantees better outcomes or demonstrates compliance. Read the Playbook’s scope.
The immediate test is practical: can the board follow a significant concern from the experience reported through management’s assessment to a decision, an action, or a clearly stated unresolved question?
For continuing research on the people, operating risks and evidence behind executive decisions, explore Touch Stone Executive Intelligence Membership.