AI in the Boardroom: Using It Well Without Giving Up Judgment
You are being told your board should use AI. You can see the upside: faster preparation, less drudgery, a sharper read on a thick pack. You can also see what keeps you up: it can be confidently wrong, it could expose confidential board material, and when it does err, you are still the one who is accountable. Most boards are stuck somewhere between the hype and the fear. The useful question is not whether to use AI. It is how to take its leverage without handing it your judgment or your duty.
Where AI genuinely helps a board
The grunt work, not the governing. Summarising a long board pack. Drafting minutes. Pulling the decisions and actions out of those minutes. Scanning the regulatory and competitive landscape and surfacing what changed. These are the time-sinks that sit between meetings and eat the hours a director should spend thinking. AI is genuinely good at volume and at first drafts. Used well, it hands those hours back.
Where it must never go
The decision. AI does not carry your duty of care, it cannot be held accountable, and it does not understand your organisation's context, your values, or the room. It also fails in a particular and dangerous way: it is confidently wrong, presenting a plausible answer with no signal that it is mistaken. A board that lets AI decide, or that takes its output as truth unchecked, has not saved time. It has quietly imported risk. The judgment, the decision, and the accountability stay with the directors. Always.
The test: a tool, not a replacement
Three questions tell you whether an AI tool is safe for a board to lean on:
- Does a human review and approve everything before it counts? Nothing should enter the record on the AI's say-so alone.
- Can you see where the information came from? A claim you cannot trace is a claim you cannot rely on.
- Is your confidential board material kept private? Privileged and in-committee matters must not be fed into a model that learns from them or leaks them.
If the answer to any of those is no, it is not a tool. It is a liability.
Getting your board's AI posture right
- Agree a short board position on AI: what it may touch, and what it may not.
- Treat every AI output as a draft to be checked, never a source of truth.
- Keep a human in the loop on anything that enters the official record.
- Guard confidentiality, especially privileged advice and in-committee material.
- Be transparent with the board about where AI is being used, and how.
How Boardside helps
We built Boardside on exactly this principle: AI as a sidekick, never the oracle. It reads your uploaded minutes and pulls out the actions, risks, and assumptions buried in them, and it scans the landscape for what matters. But nothing it produces enters your records until a person reviews it and accepts it. You always see what it suggested, where it came from, and you always make the call.
Illustrative only. You review every suggestion, and nothing is added to your records until a person accepts it.
See AI that does the heavy lifting while your board keeps the judgment, and the final say.
Book a demoThis guide is general information for New Zealand directors and is not legal advice. How your board adopts and governs AI depends on your organisation's policies, sector, and obligations. Confirm specifics with your own advisers.