Solution Story · Financial Oversight

Five million lines, re-checked every year

No finance team on earth can re-audit five million invoice lines annually. So the software does — continuously rescanning a rolling twelve months of sales history for red flags, and making sure every flag actually gets worked.

The problem

Errors that survive the first check live forever

Even with automated reconciliation and governed pricing upstream, things slip through: a price condition applied to the wrong group, a unit-of-measure mixup, a rebate missed, a margin that drifted after a cost change. Individually small; multiplied across millions of lines and a hundred branches, real money. Traditionally, these errors are found by accident — a sharp-eyed controller, an annoyed customer — or never.

Spot-checking finds the errors you were lucky enough to look at. Scanning everything finds the rest.
Why sampling isn't oversight
What we built

A watchdog with a to-do list

The module continuously rescans every invoice line from the last twelve months against red-flag rules. But detection is the easy half — what makes it work is that every flag becomes a tracked item with an owner, a status, and a deadline.

4,985,092 invoice lines scanned rolling 12 months every line, every rescan red-flag rules: pricing · UOM · margin drift · rebate misses 102,033 flagged for review (YTD) ≈2% of lines — each becomes a tracked work item 96.66% addressed on time 98,621 of 102,033 94.58% on time and closed 192,149 items closed, trailing 12 mo
The funnel: scan everything, flag the ~2% that deserve eyes, then measure the humans too — on-time resolution is a first-class KPI.
New
2,921
trailing 12 months
In progress
750
being worked now
Closed
192,149
trailing 12 months
Addressed on time
96.66%
98,621 / 102,033 YTD
On time & closed
94.58%
96,507 / 102,033 YTD

Recreated from the live production dashboard, mid-2026.

Continuous, not annual

This isn't a year-end audit project. The scan runs on a rolling twelve-month window — an error introduced today gets flagged while the invoice, the order, and the people involved are all still fresh.

Flags with deadlines

A flag that sits is a flag that didn't matter. Every item carries aging and on-time metrics, so resolution discipline is visible to management — and it shows: 96.66% on time.

Feedback for the whole loop

Patterns in the flags point upstream — a pricing condition set wrong, a vendor whose confirmations drift. The audit doesn't just fix errors; it teaches the rest of the platform where to tighten.

Results

Oversight at a scale no headcount could match

~5M
lines re-audited every year
102k
red flags surfaced (YTD)
96.66%
addressed on time
192k
items closed in twelve months
Most systems can flag problems. The rare thing is a system where 96% of the flags get resolved on time — that's not software, that's software plus accountability.
Detection is a feature; resolution is a culture