Using AI to Help Write Better Board Papers
Using AI to Help Write Better Board Papers
David Cantrick-Brooks | 21/06/2026

We all know how much time is required to write a good board paper (I am including committee papers in this) and we’ve often heard the adage: ‘I didn’t have time to write you a short letter, so I wrote you a long one instead’ (or words to that effect), often attributed to Mark Twain, although it goes back to Blaise Pascal in 1657.Regardless, both men were highlighting the same timeless truth: writing concisely is incredibly difficult. It is always faster and easier to let words spill out onto the page; editing them down into a sharp, clear, and powerful message requires significant time and effort, especially where there is a lot of data and the issues are complex.Fortunately, we have AI to call upon to help us, although we need to be very careful about confidentiality and sensitivity – internally and externally.There must be appropriate guardrails (often in the form of policies and procedures, etc.) and they need to be followed by everyone (without exception).

Directors want papers which (among other things):

You will note that the word ‘relevant’ appears frequently, which should come as no surprise.Problems with paper length can often be traced back to a lack of relevance in the content.

As a wise mentor once said to me, when writing a meeting paper always think about the WHO, WHAT, WHEN, WHERE, WHY and HOW, noting there can be many angles to consider here.

So, how can AI help with these?With well-constructed or engineered prompts, generative AI (combined with agentic AI1) can:

Diarise due dates for papers and send reminders to management.Think workflow …

Research and compile content from relevant, valid and credible internal and external sources.

Succinctly summarise and synthesise large volumes of data / information, even if it is dispersed and complex.

Construct tables, graphs, flowcharts, etc. with minimal effort.

Identify, describe and explain relationships between things.

Test assumptions, limitations and constraints under certain scenarios, including with respect to different alternatives.However, generative AI can produce inaccurate or fabricated content; the Federal Court’s 2026 AI practice note is a useful reminder that outputs may contain fictitious citations, factual errors or incorrect legal analysis.

Reference previous relevant decisions / transactions.

Populate standard templates with the required information.

Review papers for completeness, logic and accuracy (including with reference to any agreed guidelines / instructions to authors of papers or any specific requirements).

Edit papers according to certain guidelines (including tone and style).

Call out issues requiring further thought / attention.

Tidy up papers ready for publication.

Create handy summaries for directors.

Identify emerging themes and issues within a meeting pack for directors, which might otherwise be missed.

Create draft ‘skeleton’ or ‘shell’ minutes and run sheet from the papers in the meeting pack for the company secretary.

Assign responsibilities in respect of pre-approval sign-offs and post-approval actions.Again, think workflow …

Send reminders to those responsible for post-meeting actions.

Provide inputs into the meeting cycle (including, for example, additions to the annual workplan).

IMPORTANT: This list is not exhaustive (simply illustrative).

1 Note that AI can diarise, assign actions and set and send reminders only where it is properly integrated with authorised systems, configured correctly and subject to access controls.

It can take time to construct or engineer an AI prompt, but the time and effort will be worth it.I can say this from first-hand experience.

AI is an incredible tool, although we need to be very careful about confidentiality and sensitivity – even with an enterprise version.There must be appropriate guardrails (in the form of policies and procedures, etc.) and they need to be followed by everyone.Note that approved enterprise tools also require appropriate data-handling rules, user training, human review and governance controls – consistent with the Australian Government’s voluntary AI safety guardrails, which emphasise accountability, risk assessment, testing, controls and meaningful human oversight.

OAIC guidance recommends that organisations do not enter personal information, particularly sensitive information, into publicly available generative AI tools because of significant and complex privacy risks.

With deep knowledge and experience, Governance in Action can assist clients with ‘traditional’ and AI-assisted board paper writing, including prompt engineering.Governance in Action also has and can develop bespoke AI-based apps to assist clients with writing board papers.

This article has been prepared with the aid of AI. Final judgement, editing and accountability of course remain with the author.

PreviousNext

Related Articles

Local AI Models: A Governance Opportunity, Not a Governance Shortcut

Local AI models may allow organisations to use generative AI with sensitive information that would be unsuitable for consumer tools and difficult to approve for some cloud-based services. For boards, executives and company secretaries, that could be a significant opportunity. But local AI is not a governance shortcut. It still requires clear rules for data classification, privacy, records retention, legal holds, model provenance, cyber security, human verification and board reporting. The question is no longer simply whether an organisation should use AI. It is which AI should be used, for which data, and under which controls.

06/21/2026

The Digital Director: Boardroom Copilot, Not Licensed Driver

AI-enabled boardroom assistants are moving quickly from novelty to practical governance infrastructure. These tools can summarise board packs, search historic minutes and resolutions, draft questions, prepare minutes, track actions and surface emerging risks. Some vendors now use bold language such as "AI Board Member". Yet, under Australian corporate law, AI is not a director. It cannot vote, owe fiduciary duties or replace independent human judgement. This article considers whether boards are ready to use AI in the boardroom, where the legal and practical boundaries lie, and what guardrails directors, company secretaries and general counsel should apply. The conclusion is deliberately balanced: boards should not ignore the opportunity, but nor should they let AI drive. Controlled pilots, approved tools, source verification, human review, data-handling controls, cyber due diligence and clear board policies are essential.

06/20/2026