Automating an Eight-Step Monthly Group Reporting Pack
Sector:
South African subsidiary of a multinational industrial manufacturing group.
Key outtakes: An eight-step monthly reporting pack — covering management accounts mapped to more than 90 group cost elements, more than 2,000 inventory line items, a 12-month rolling sales analysis across more than 50 customers and over R300 million of revenue, and a revenue data warehouse upload of almost 3,000 invoices and more than 29,000 order lines — produced as a single automated workflow every cycle, with every figure traceable from raw source data through to final submission.
The Challenge
The client is the South African subsidiary of a multinational industrial group, and is required to submit a comprehensive monthly reporting pack to the group head office in a tightly defined format. The pack consists of more than eight distinct uploads each month covering management accounts, year-to-date consolidation, inventory, sales, selling, general and administrative costs, accounts receivable credit risk, and a full revenue data warehouse.
The underlying data sits across two source systems — a local accounting platform and a separate ERP. The local chart of accounts, cost centres, business units and cost elements all need to be translated into the group’s reporting structure, and each component of the pack needs to reconcile back to the local books before being submitted upstream.
The client needed the entire reporting pack produced reliably, in the group’s format, every month, with a clean audit trail from the source systems through to the final submitted numbers — and without the recurring management cost of pulling the local finance team off operational work for a week each month.
Why Conventional Methods Fell Short
The pack is not one report; it is eight, each with its own mapping logic, its own data sources and its own reconciliation back to the local trial balance. Volumes are non-trivial: more than 2,000 inventory line items, almost 3,000 invoices and tens of thousands of order lines per cycle, a full management accounts mapping into more than 90 group cost elements across multiple business units, and selling, general and administrative cost mappings across more than 20 cost centres.
Manually, this is a one-week-per-month exercise that has historically required senior finance involvement, and remains exposed to transcription error, inconsistent month-on-month formatting, and resubmissions where the group rejects an upload that does not tie out.
The historical workaround — at this client, and at most subsidiaries with the same problem — was a stack of intermediate Excel files, where data extracted from the source systems was massaged into shape before being uploaded. Those intermediate files drift in format month to month, the mapping logic embedded in their formulas is undocumented, and the final figures often cannot be traced back from the upload through to the source data because numbers get hardcoded along the way. That structure is fragile by design, and any automation built on top of it inherits the fragility.
The other constraint is cadence. This is not an annual procedure. It runs every month, twelve cycles a year, every year — which means the cost of any manual inefficiency is paid twelve times over, and the value of automation compounds in the same way.
Our Approach with Monsoon
We rebuilt the reporting pack as an eight-step Monsoon workflow running directly off the raw source data — the local accounting platform exports, the ERP data, and the trial balance — rather than off the intermediate files the client had historically used. Each step in the pack is a separate Monsoon agent with a documented, traceable transformation from raw data to final upload format.
Group-level values that previously sat as hardcoded outputs are now generated by matching the raw data directly against the group’s accounting manual, so every figure is traceable back to a source document rather than to an undocumented spreadsheet formula. Validation and reconciliation checks are built into each stage of the workflow, and a monthly reconciliation report is produced automatically as part of the pack rather than performed as a separate manual procedure after the fact.
The shift from intermediate files to raw data was the methodological turning point on this engagement — and it is a transferable principle that now applies to every Monsoon automation we build: the agent runs against the authoritative source, not against an artefact of someone else’s manual process.
Outcomes for the Client
- The full monthly reporting pack produced as a single automated workflow , every cycle, in the group’s required format.
- Every figure traceable from raw source data through to final submission — for the first time, a fully reproducible audit trail from source to upload.
- A week of senior local finance time returned to the business, every month — capacity that previously went into manual extraction, mapping and reconciliation now moves to actual management of the business.
- Resubmissions to group fall away , because the workflow produces a consistent, validated output every cycle rather than a format that drifts month to month.
What This Meant for RAiN
This is a recurring monthly engagement rather than a one-off annual procedure, with the value compounding across each cycle. The methodology is directly reusable on any subsidiary of any multinational with a similar group reporting obligation — a large addressable population of clients. The engagement also extends Monsoon’s commercial story beyond audit and assurance into operational finance automation, broadening the platform’s market without requiring fundamentally different infrastructure.
The methodological lesson learned on this engagement — build off raw data, not intermediate files — has become a standing principle in how every subsequent Monsoon workflow is designed.