Almadar A board that meets in a KAFD tower still has someone manually assembling data from three systems the night before, the exact work AI handles best.
A monthly or quarterly board pack in a Riyadh group typically consumes days of finance time in collection, formatting and drafting, most of which is mechanical. Automating that assembly frees the finance team for the analysis that actually adds value, and shortens the gap between period end and the board being informed.
Automated assembly and drafting
Where data sources are connected, the pack can be generated: financial statements, KPI dashboards, variance tables and segment analysis populated automatically. AI can then draft commentary explaining what moved and why, drawing on the underlying data, which finance reviews and corrects rather than writing from a blank page. The draft is a starting point, not an output.
Bilingual reporting
For Saudi boards this is a genuine and practical gain. Producing Arabic and English versions of the same pack has historically meant either duplicated effort or translation with the risk that a late adjustment updates one and not the other. Generating both from the same data removes that risk, and connects to Arabic financial reporting configuration.
What must remain human
Deciding what the board needs to know is judgment. So is any commentary on outlook, risk or management intent, none of which a model can source from historical data. And nothing goes to a board without the numbers being verified against source, because a plausible but wrong figure in a board pack is a governance failure regardless of how it was produced.
A common Saudi scenario
A Riyadh group's board pack takes seven working days to assemble across four entities and two languages, meaning the board reviews figures three weeks after period end. Automating assembly and first-draft commentary brings this to two days. The board discusses the quarter while it is still current, and the CFO spends the recovered time on the analysis the board actually asks about.
Design before automation
Automating a board pack nobody reads carefully simply produces the wrong thing faster. We usually review what the pack contains against what the board actually discusses first, which frequently removes a third of the content, and only then automate what remains. This is a reporting design exercise as much as a technology one, and it draws on the same data foundations as executive dashboards.
Version control and the audit trail
Board packs get revised, sometimes late and sometimes after distribution. Automated generation must handle versioning explicitly: which version went to the board, what changed between versions, and who approved the final. Without this, a board can end up discussing different numbers in the same meeting, which is a governance problem rather than an inconvenience. The same discipline underpins consolidation and connects to governance over automated outputs used in formal decision-making.
Family-owned groups in Riyadh with boards including non-executive family members benefit most from bilingual generation and clear plain-language commentary, while listed and pre-IPO groups in Riyadh prioritize speed to meet regulatory reporting timelines.