Drafting the Board Pack With AI

The board pack — the collection of financial and other information prepared for the board — is an important and often time-consuming deliverable of the finance function, and one where AI can genuinely help with the drafting, particularly the narrative and explanatory elements. AI can assist in drafting parts of the board pack — the commentary, the summaries, the explanations — capturing a time saving on a task that involves considerable writing. But the board pack is a high-stakes deliverable, presenting the financial position and performance to the board for their governance and decisions, which means it must be accurate, appropriate and sound, and AI’s assistance must be used carefully, with the finance professional ensuring the board pack meets the standard the board requires. This guide addresses drafting the board pack with AI, capturing the value while ensuring the board pack is accurate and appropriate.

This guide is written for Financial Controllers, Finance Managers and the accountants who prepare board papers. It covers where AI can help, why the board pack’s high stakes require care, how to use AI well for the drafting (including inputs and example prompts), the finance professional’s essential role in review and sign-off, and the confidentiality rules and controls that should sit around it. It builds on the general discipline covered in our guide to board pack preparation and on what the board actually reads, with a focus on using AI for the drafting.

Where AI Can Help With the Board Pack

AI can genuinely help with the parts of the board pack that involve writing and suit its capabilities. It can help with the narrative and commentary — the written explanations of the results, the commentary on performance, the narrative around the numbers — which involve considerable writing and suit AI’s strength with text, providing first drafts the finance professional then refines and verifies. It can help with summarising — producing summaries of the results, the position or particular matters for the board — as a starting point. And it can assist with explaining — articulating explanations of matters for the board in clear prose.

AI can also help with the more routine, recurring narrative elements of the board pack that follow a consistent pattern month to month, producing them in the established style and format for the finance professional to refine. It is also useful for checking that figures quoted in the narrative match the tables and that commentary is consistent with last month’s messages, and for outlining a paper before any figures are added.

A finance professional who uses AI for these tasks, then refines and verifies the output, can save real time on the board pack — but it is not a substitute for deciding what the board needs to know. Our wider guide to how a Financial Controller uses AI sets the board pack alongside the other controlling tasks where AI helps.

Why the Board Pack’s High Stakes Require Care

The board pack is a high-stakes deliverable, and its stakes require care in using AI for it. It presents the financial position and performance to the board — the body responsible for governing the business — for their oversight and their decisions, so it informs the governance of the business at the highest level. Directors have legal duties and, in larger and listed companies, governance expectations such as those in the UK Corporate Governance Code that depend on receiving accurate, timely information. Errors or unsound material in the pack could mislead the board and impair its governance, with potentially serious consequences.

This means using AI for the board pack must not compromise its accuracy, appropriateness or soundness. AI language models can produce output that is fluent, confident and wrong; they can invent a plausible cause for a variance, introduce a percentage they have miscalculated, or quietly soften a caveat because the surrounding text is positive. Anything AI contributes to the board pack must therefore be carefully verified by the finance professional, not trusted uncritically. The board pack’s importance to the board’s governance means its integrity cannot be compromised by careless reliance on AI.

There is a second kind of stake: confidentiality. Board papers are among the most sensitive documents a business produces, so where they go when pasted into an AI tool matters as much as what comes back — see the confidentiality section below.

How to Use AI Well for the Drafting

Using AI well for drafting the board pack means using it to assist with the writing while the finance professional ensures the board pack is accurate, appropriate and sound. The effective approach is to use AI to draft the narrative, commentary and explanatory elements while the finance professional provides or verifies the substance, so the content accurately reflects the actual position and performance. AI helps with the writing; it does not determine the substance or produce content that goes to the board unverified.

In practice that means giving AI the genuine information and context, using it to articulate them clearly, then refining and verifying the drafted content before it goes into the pack. The finance professional must be particularly careful with anything factual or consequential. The same principles apply to the monthly management pack, covered in our guide to AI for management reporting narrative; board papers add a higher bar for accuracy and confidentiality.

What to give the AI

The quality of the draft depends on the quality of the inputs. Before opening any tool, prepare:

• The figures, already calculated and checked, from the finance system or reporting model. AI tools are not reliable calculators, so the tool should compute nothing new.
• Your explanations in your own words: the real cause of each material item, timing or permanent, and any action agreed.
• The board’s concerns this period, so the draft addresses what directors are actually watching.
• The house style: last period’s approved paper, plus a materiality threshold and word limit.

What you leave out matters too: strip out anything that identifies an individual, and keep transaction details, inside information and individual remuneration out of the tool unless policy and legal review allow it.

A prompt structure that works

A reliable prompt has six parts: role and audience; task and word limit; the data, with variances already calculated; your explanations as the only permitted causes; constraints (calculate nothing, invent nothing, flag gaps); and format and tone (conclusion first, plain English, UK spelling). Telling the tool to mark gaps with a placeholder turns its weakness into a check: any placeholder left in the draft is a question the finance team still has to answer. The underlying skill is covered in prompt engineering for finance.

Example prompts with placeholders

Replace each item in square brackets. The first contains no figures; the others assume an enterprise tool approved for board papers and figures checked before they are pasted. They are consistent with the board prompts in our finance prompt library.

1. Section outline (no figures). “I am preparing the finance section of a board pack for [TYPE OF BUSINESS] for [PERIOD]. The board’s main concerns this period are [CONCERNS]. Propose a structure of no more than [NUMBER] headings, with one sentence on what each should cover and the key question the board is likely to ask under each. Do not include any figures.” Outlining first and adding figures yourself keeps sensitive numbers out of the tool.

2. Finance report first draft. “You are helping the Financial Controller of a [SECTOR] business draft the finance report for the [PERIOD] board meeting. Using only the figures and notes below, write no more than [WORD COUNT] words. Open with the single most important message. Cover performance against budget, cash and [NUMBER] matters the board needs to know. Do not calculate any new figures or add causes not in my notes. Where a variance above [MATERIALITY THRESHOLD] has no explanation, write [EXPLANATION NEEDED]. Figures: [PASTE TABLE]. Notes: [PASTE NOTES].”

3. Tightening a draft. “Edit the draft below for a board audience. Make it shorter by about [PERCENTAGE], lead with the conclusion, and remove jargon. Do not change, add or remove any numbers, and do not change the meaning. Mark any sentence you think is unclear with [CHECK]. Draft: [PASTE DRAFT].” Often the most useful, lowest-risk use.

4. Consistency check. “Compare the narrative below with the tables that follow. List every figure in the narrative that does not appear in the tables or differs from them, and every statement that conflicts with last period’s commentary. Do not correct anything; list the issues only. Narrative: [PASTE]. Tables: [PASTE]. Last period: [PASTE].”

The Finance Professional’s Essential Role

The finance professional’s role is essential, because ensuring the board pack is accurate, appropriate and sound requires understanding, judgement and verification that AI cannot provide. The finance professional must ensure the pack accurately reflects the actual financial position and performance, which requires their understanding of the substance, not the tool’s plausible generation. They must ensure the content is appropriate for the board — pitched rightly, presenting the right matters in the right way — which requires judgement about what the board needs. And they must verify that the pack is sound, taking ownership of it as the deliverable that goes to the board.

AI assists with the drafting but does not replace any of this: “the AI wrote it” is not an answer to a director’s question. A finance professional who trusts AI’s output uncritically for a high-stakes deliverable risks a board pack that misleads or fails the board. The role cannot be delegated, which is why our guide Can AI replace a Financial Controller? concludes that accountability keeps the role firmly human.

Review and sign-off

Review AI-drafted board content in a fixed order, so nothing is skipped under deadline pressure:

• Figures. Trace every number to the tables; reject any figure the tool calculated itself.
• Causes. Delete any explanation you did not supply, however reasonable it sounds.
• Completeness. Every material item and board concern covered; every placeholder resolved.
• Caveats. Risks, uncertainties and estimates present and not softened.
• Consistency. The paper agrees with the tables and other papers, and explains any change of view.
• Sign-off. A named reviewer approves each AI-assisted section before issue.

Our guide to human-in-the-loop AI controls shows how to build this review into the reporting timetable rather than bolting it on at the end.

How to Approach Using AI for the Board Pack Sensibly

A finance professional should approach using AI for the board pack sensibly: using it to help draft the narrative, commentary and explanatory elements while ensuring the substance is genuine, the content appropriate and the pack sound — and verifying thoroughly, because the board relies on the pack and the consequences of errors are serious.

It also means using AI as a drafting aid that expresses the genuine substance, not as a source of the substance, and the finance professional retaining ownership of the pack and its quality. A practical way to start is with the lowest-risk use — editing a paper a person has already written — then moving to first drafts from notes once the review routine is working. For the first cycle, run an AI draft alongside the normal manual draft and compare them before relying on the AI version. This connects to the guidance on board pack preparation and the safe-use principles in where AI helps and where it is dangerous.

Maintaining the Board Pack’s Voice and Judgement

A particular consideration in using AI to draft the board pack is maintaining its appropriate voice and the finance professional’s judgement, because the board pack is not just information but a considered communication to the board. It conveys the finance function’s view of the position and performance, pitched for the board, reflecting judgement about what matters and how to present it — and that considered voice is part of what makes the pack valuable. AI-drafted content, used uncritically, could produce a pack that is generic or that lacks the judgement it should carry. AI tools also tend towards an upbeat style that can soften bad news.

Maintaining the voice and judgement means the finance professional shaping the pack to reflect their considered view, using AI to help articulate it rather than determine it. Give the tool a short style note (conclusion first, bad news stated plainly, no adjectives without evidence) and put the “so what” for each material item into your notes, so the draft leads with implications rather than describing the table. A useful test before sign-off: if a sentence could appear unchanged in any company’s board pack, rewrite it or remove it. Our guide to variance analysis that drives decisions covers how to find that insight in the first place.

Confidentiality: Board Papers Belong in Enterprise Tools Only

Board papers are highly sensitive. Under the green, amber and red classification set out in our guide to AI data security in finance, board papers and management accounts are amber: they belong only in an enterprise-tier tool the business has approved and covered with a data processing agreement, never in a free or personal consumer account. Consumer tools may retain inputs and use them to train models; enterprise plans typically exclude customer data from training under contract.

Some board content is red and should not go into an AI tool without legal review: personal data such as individual salaries or disciplinary matters, M&A and funding information subject to confidentiality obligations, regulatory submissions and, for listed companies, inside information. Confidentiality is also a fundamental principle of the ICAEW Code of Ethics, and where personal data is involved the ICO’s guidance on AI and data protection is clear that the organisation remains responsible however the tool is provided. With no approved enterprise tool, use only the figure-free outline prompt.

Controls and Audit Trail

Because the board pack feeds the highest-level decisions in the business, the controls around AI use should be written down rather than held in one person’s head. Most belong in the team’s AI usage policy; for board papers specifically, the essentials are:

• Approved tools only. Board material goes into enterprise tools the business has approved, accessed through company sign-on, never personal accounts.
• Standard prompts. Agreed prompts from the prompt library, so the constraints are applied every time and the method is easy to explain.
• Named reviewer. Every AI-assisted section has a named reviewer who signs it off before the pack is issued.
• Records. For each paper, a note of the tool used, the prompt or a summary of it, the source data and the reviewer, filed with the working papers.

Directors and auditors increasingly ask how AI is used; a simple record answers quickly. Our guides to what boards are asking about AI in finance and what auditors ask about AI cover what to expect.

Frequently Asked Questions

Can AI write the board pack on its own?

No. It can draft narrative from figures and notes the finance team supplies, but it does not know why the numbers moved or what the board needs to hear. Without those notes it will fill gaps with plausible causes. A qualified person should review and approve every AI-assisted section before the pack is issued.

Can we paste board papers into ChatGPT?

Not into a free or personal account. Board papers should only go into an enterprise-tier tool your organisation has approved, with a contract covering data use and training. Personal data, transaction information and inside information need legal review before any AI use.

Should we tell the board that AI helped draft the pack?

Be ready to explain it and follow your AI usage policy on disclosure. Accountability stays with the finance team whichever tool helped.

What skills should we look for in someone preparing the board pack?

Strong reporting, judgement and clear writing first; then the ability to prompt with constraints and spot plausible-but-wrong output. Our Financial Controller recruitment service assesses these capabilities, and our guide to how AI is changing what employers look for in an FC explains how employers are testing them.