Automating board reporting with Power BI: replacing the monthly pack

The board pack is the most manual document most finance teams produce. Somewhere between the third and the tenth working day of the month, one or two people export the trial balance, refresh a workbook that has grown for years, paste the outputs into slides, write commentary, and send a PDF that is out of date by the time the board reads it. When a director asks a question, the answer is another export and another day.

Automating board reporting means the numbers in the pack come from a model that refreshes itself, the definitions cannot drift between months, and a director's question can be answered by clicking rather than by exporting. It does not mean the pack disappears. This article sets out what changes, what stays, and the order in which to do it.

What "automated" actually means

A board pack is automated when four things are true. The figures come from a governed Power BI model that reads the finance system on a schedule rather than from an export. Every measure in the pack has one written definition that is the same in the pack as it is in the daily reporting. The comparison periods, budget versions and consolidation rules are set in the model, not chosen by hand each month. And the pack can be regenerated for any month, including a prior one, without anyone rebuilding it.

Anything less than that is a faster export, which is worth having but is not the same thing. The test is whether the pack for last month can be reproduced today, unchanged, without touching a workbook.

Lock the definitions first

The reason board packs are hard to automate is rarely technical. It is that the pack has accumulated definitions nobody agreed to. Gross margin excludes freight in one table and includes it in another because two people built the two tables in different years. Headcount is FTE on one page and heads on the next. Revenue is invoiced on the P&L page and recognised on the KPI page.

The first piece of work is therefore a definitions review: every measure that appears in the pack, listed, with its formula, its source, its comparison basis and its owner. This is usually a two-week exercise with the finance team and it produces the most valuable document in the project, because the board then signs off the definitions once and the arguments in the meeting stop being about whose number is right. Only after that is anything built.

Build the model, then the pages

The model is the part the board never sees and the part that determines whether the pack is trusted. It reads the ledger lines from the finance system, the budget from wherever the budget lives, the payroll summary, and whichever operational sources the pack draws on: sales by channel from the POS or the ERP, occupancy from the property management system, tonnes moved from the warehouse system. Each source lands in a table with a documented refresh; each measure is written once against those tables; and the entity, currency and period logic is set in the model rather than in a visual.

The pages come after the model, and they should mirror the pack the board already knows: the same order, the same names, the same measures. Changing the layout at the same time as changing the source doubles the number of things the board has to learn to trust. The redesign, if one is wanted, comes in a later quarter, when the numbers have earned belief.

The narrative layer

Directors read the commentary before the charts, and the commentary is the part of the pack that cannot be generated. What can be automated is everything around it. A Power BI page can carry a text box per section that the CFO or controller writes each month against the refreshed figures, with the figures themselves pulled into the sentence so that "revenue was 6% under budget" updates when the model refreshes and only the explanation is typed. Exceptions can be surfaced automatically: the variances beyond a threshold, the debtors past a days limit, the reconciliations still open, so the writer starts from a list of what needs explaining rather than from a blank page.

The commentary is also the reason the pack should remain a document as well as a report. A board minute refers to a pack as it stood on a date; a report that has refreshed since cannot be that record. The answer is to export the pack from Power BI each month, as a PDF or a PowerPoint, and keep the export as the minuted version while the live report stays available for questions.

What changes in the meeting

The visible change is that the meeting stops opening with a reconciliation. When the figures on the screen are the same figures the finance team and the operations team have been looking at all month, and the definitions were agreed at the start, the first twenty minutes are spent on what to do rather than on which number is right. Questions that used to become actions for the next meeting are answered in the room, because the director can drill from the summary to the customer, the site or the cost centre while everyone is looking at it.

The less visible change is in the week before the meeting. The finance team is not assembling; it is reviewing. That is a different job, and a better use of a controller's time, but it is worth being honest that it is still a job. Automating the pack removes the manual assembly; it does not remove the need for someone to read the numbers and decide what they mean.

Multi-entity groups and the UAE specifics

Most boards in Dubai and the UAE oversee more than one entity, and the pack has to consolidate a mainland company with a free-zone company and sometimes a foreign subsidiary on a different system. That consolidation is where hand-built packs spend most of their manual hours: mapping each entity's accounts to a group chart, eliminating intercompany, translating currency. In the model, each of those is a rule applied on refresh. The VAT position by entity and the split of income between entities can be shown on the pack as figures reconciled to the ledger; whether those figures are right for a return or a Corporate Tax computation is a question for the tax agent, and the pack should say so rather than imply otherwise.

Where to start

The definitions review is the right first step whatever the eventual tooling, and it is short enough to be scoped and quoted on its own. If the pack is still in Excel and the question is whether to automate it there or move it, the Excel to Power BI migration page sets out how that decision is made. If the model already exists and the pack is the missing piece, the Power BI dashboard development page describes what a fixed-price build of the pack pages includes.