Skip to main content

    AI strategy and governance

    How to brief your board on AI

    Board AI briefings need clarity, commercial framing and open discussion of risk. A good briefing connects each AI initiative to revenue, margin or time saved and outlines the governance around it. It also addresses the Privacy Act and sets out risks such as bias, security and regulatory exposure. This article gives executives a structured approach for presenting AI proposals to Australian boards.

    In short

    Open board AI briefings with the effect on profit and loss

    • Board AI briefings often fail because they lead with technology. Open with the effect on profit and loss (P&L), and leave the model architecture for questions.
    • The Five Principles of Strategic AI Briefing give executives a repeatable structure that directors engage with.
    • Pair every AI opportunity with its risk profile from the Board AI Risk Register.
    • Pilots that start without a baseline, an owner and a decision rule are less likely to reach production. Your briefing must show why this one will, with stop criteria in case it does not.

    Detail

    Five briefing principles, led by commercial consequence

    Many board AI briefings lose their audience early. The executive opens with an explanation of large language models, neural networks or fine-tuning. Directors disengage, and by the time the business case arrives, their attention has gone.

    The Five Principles of Strategic AI Briefing fix this.

    The first principle is to lead with commercial consequence. Open with the dollar figure, because a sentence stating the annual margin improvement you are targeting gets attention. A description of the model's architecture does not.

    The second principle is to frame risk alongside opportunity. Pull the relevant entries from your Board AI Risk Register and present them together. Directors approve initiatives they can assess. If you show only the upside, they will assume the work has not been done.

    The third principle is to compare with competitors. Show where peers and competitors are deploying AI in the same area. Evidence of competitors moving ahead focuses a board's attention.

    The fourth principle is to define success and failure criteria in advance. Your board will know that pilots without a baseline, an owner and a decision rule often stop making progress. Address this directly. Set out the measures that prove the initiative works, the criteria for stopping it and the most capital at risk before a decision to proceed.

    The fifth principle is to present the request with options. Give the board 2 or 3 options with different risk and investment profiles, and let directors choose.

    Keep the briefing short, and allow at least as much time for discussion as for presentation. A proposal is ready when you can explain it in that time.

    The Board AI Risk Register helps in another way. It gives directors a familiar framework for assessing AI proposals, because they have seen risk registers for cyber, compliance and operational hazards. Presenting AI risk in the same format makes it easier to assess. Each well-structured briefing also builds the board's AI literacy over time.

    Commercial impact

    Board alignment decides whether AI initiatives proceed

    Board alignment decides whether AI initiatives are funded, scaled or stopped. A poor briefing wastes meeting time, and it can set your AI program back by months.

    Directors need enough information to oversee AI. The Australian Securities and Investments Commission (ASIC) published Report 798 in October 2024. It states that directors' and officers' duties of reasonable care and diligence 'extend to the adoption, deployment and use of AI'. A good briefing seeks approval, and it also gives directors what they need to meet those duties.

    Poor briefings have lasting effects. Boards that do not understand AI tend to over-restrict, stopping every proposal out of caution, or under-govern, approving everything without scrutiny. Both outcomes damage the organisation.

    Shadow AI is unapproved AI use, and it reaches the executive team as well as junior staff. A board briefing that does not acknowledge shadow AI and offer a governed alternative leaves an obvious risk unaddressed.

    Boards that receive effective AI briefings can also make faster, better-informed investment decisions.

    If a regulator asks your board about AI oversight, the quality of your briefing papers becomes evidence of how well the board governs AI. Good briefings also build the board's AI literacy over time, so each later proposal moves faster and draws sharper questions.

    In practice

    Illustrative cases of how board AI briefings open

    Illustrative composite: a chief financial officer at a mid-sized logistics company opens her AI briefing with the annual savings expected from automated freight routing. She leaves the optimisation algorithm for questions. She maps 3 risks from the risk register and shows how far competitors have adopted similar tools. She defines a time-boxed pilot with monthly measures, then presents 3 investment tiers for the board to choose from.

    Illustrative composite: a chief information officer opens a generative AI briefing with the problem of unapproved AI use. Staff are already using ChatGPT with company data and no controls. He presents the risk, the governed alternative and the cost of doing nothing, so the board has a specific decision to make.

    Illustrative composite: a chief executive presents an AI customer analytics tool with a long technical walkthrough. The board asks for a plain English version and defers the decision to its next meeting.

    Illustrative composite: another board uses a standard AI briefing template. The first paragraph covers commercial impact, the second the risk register entries and the third the competitive context. The fourth sets out success and stop criteria, and the fifth makes the investment request with options.

    What to do

    Steps to structure a board AI briefing for a decision

    • Apply the Five Principles of Strategic AI Briefing: commercial consequence first, risk beside opportunity, competitor comparison, success and stop criteria, and a request with options.
    • Keep briefings short and allow at least as much time for discussion as for presentation. A proposal is ready when you can summarise it briefly.
    • Include the relevant Board AI Risk Register entries with every proposal.
    • Deal directly with the risk that pilots stall before production. Show your stop criteria and the most capital at risk.
    • Standardise your AI briefing template so the board can decide in the meeting.

    Last reviewed

    Put this topic against your own numbers, owners and timelines.