Five source files, five real results — every one built from synthetic data, none of it a client's. Download the source and see the numbers. Run your own file to see your aggregate total.
Every sample below is generated data, not a real customer's file. Each one is picked to prove something specific — not just "here's a report", but why this particular file looks the way it does.
Two different things. Every sample report below is open in full — all three documents, every finding, no sign-up. They're synthetic, so nothing needs hiding. Your own preview is a demonstration: it shows your file's aggregate total — how many Duplicate-Payment Groups it found and the dollars they add up to — not the itemized findings. The full, line-by-line findings on your file come with the full report.
01 · The honesty check
A genuinely clean file.
No duplicate payments, no vendor spelling variants, no missing cells, no sequence gaps — on purpose. If we invented problems to sell you a report, this is the file that would prove it. It doesn't.
3,254 rows of ordinary AP mess — case-variant vendor names, a handful of duplicate payments, a few sequence gaps, some missing cells. Not a stress test, not adversarial. This is what shows up.
Arabic, Thai, Turkish, German, and English ledgers — different scripts, currencies, and date conventions — uploaded together as one folder. The structural pass reads every one without breaking. (Money-in-dollars findings today only run on the English-headers file inside it — extending that to other languages is next.)
The same 495-row file behind the sample findings on our homepage — taken all the way through: source file in, Review out, and the cleaned file with a full change log back, exactly what a client receives when a cleaned file is requested.
GOLDEN POD