The 2026 AICD and Human Technology Institute guide recommends that boards and management agree the indicators and reporting cadence they expect for AI. It says reporting should show where AI is used, how systems are performing, where risk is changing and whether use remains aligned with strategy and risk appetite.
Report the balance between opportunity and risk
AI reporting can become a risk register with no view of value, or a list of innovation projects with no view of governance. The board needs both.
A good report shows how AI supports organisational objectives, how the portfolio is progressing, what value has been realised, and where management needs to strengthen controls or reconsider a decision.
Use a stable report structure
- Purpose and coverage. State the reporting period, accountable executive, business-unit coverage and source records used.
- Strategy progress. Show objectives, roadmap milestones, priority initiatives and agreed measures of progress and value.
- AI portfolio. Summarise active, proposed, paused and retired uses, including material additions since the previous report.
- Governance status. Show assessment and approval status, higher-impact uses, open conditions, overdue reviews and evidence gaps.
- Incidents and change. Explain material incidents, supplier or model changes, changed uses and the effect on prior decisions.
- Management action and board decisions. Name important actions, owners and dates, then state each decision or risk-appetite question clearly.
Choose measures that explain progress
Measures should be few enough to understand and stable enough to compare over time. A practical set may include:
- Visibility: business-unit response coverage, registered uses and newly discovered uses.
- Strategy: objectives on track, roadmap milestones, use cases linked to priorities and initiatives needing a decision.
- Governance: assessed, approved, conditional and unapproved uses; overdue reviews; open material actions.
- Change: incidents, material changes, supplier changes and uses returned to assessment.
- Value: expected and realised benefits for use cases where evidence is reliable, with costs shown on a comparable basis.
- Human impact: material workforce, customer or stakeholder effects and how management is responding.
The AICD and HTI recommend specific, measurable use cases when assessing AI returns. Broad claims that AI has improved productivity are less useful than an agreed baseline and a measured change in a defined process.
Explain movement, not only totals
An increase in registered uses may mean that AI adoption is accelerating, or that discovery has improved. A reduction in high-risk uses may reflect better controls, retirement of a use, or a change in classification. Give directors the reason behind material movement.
Put exceptions in context
A single score rarely tells the board what to do. For each material exception, explain the purpose of the use, why it matters, the control position, management's judgement and whether the decision remains within risk appetite.
State the basis of management's view
A credible report says what the portfolio view is based on. This may include completed discovery responses, the current AI register, assessments, supplier evidence, incident records and owner reviews. It should also identify gaps, such as business units that have not responded or supplier information that remains unknown.
This report is based on the organisation's current AI register, completed business-unit discovery responses and evidence recorded at the reporting date. The report identifies areas where coverage or supplier information remains incomplete.
Create an operating rhythm
Board reporting works when the underlying records are maintained between meetings. Owners update use cases, controls and measures as work progresses. Incidents and material changes enter the same workflow. The board pack then becomes a view of current management information rather than a quarterly reconstruction.
Sources and further reading
- AICD and Human Technology Institute, A Director's Guide to AI Governance, Version 2
- Governance Institute of Australia, AI is in the boardroom. Its governance isn't yet
- National AI Centre, Guidance for AI adoption: implementation guidance
- ASIC Report 798, Beware the gap
This article provides general governance guidance. Board reporting should be adapted to the organisation's strategy, sector, board calendar and material risks.