The team was tracking the right things. Paid on Time, OCR, automated invoice processing, PO Compliance. The metrics made sense and the reports reflected what the team actually needed to manage.
The problem wasn't what was being tracked. The problem was how data got into the reports.
Once a month someone downloaded CSV files, saved them to a shared folder, and Power BI read from there. It sort of worked. Until it didn't.
An extra comment, a comma in the wrong place, or a minor change in CSV structure and the whole processing broke. Finding exactly where the issue came from could take a week. No one had time for it, but it also couldn't be left alone, so the same problem kept coming back.
“The report is broken again. Why is this whole thing so slow and so fragile?”
Instead, we set up an automated process that picks up exports directly from scheduled emails and passes them reliably into the pipeline.
Structural differences, worse formatting, or inconsistencies between exports are handled before data reaches the reports. Instead of fragile CSV files on a shared folder: stable tables to build on.
The fragile parts are gone. No one spends days or weeks hunting down broken CSVs. The process runs stable and without outages.
In the manual steps, the fragile exports, the infrastructure dependent on one person. That's where most of the delays and most of the errors come from.
Nothing complex was needed here. Just take what the company already had, automate the entry point, and remove the manual interventions. Since then it runs on its own.