Invoice reconciliation that shows its work.
Drop in the invoices and the bank statement. InvoiceMatch scores every possible pairing across four weighted dimensions — invoice number, vendor, amount, date — and hands you the matches with the arithmetic still attached. Expand any row and you see exactly why it scored what it did. What's left is a short queue of genuine exceptions, not a spreadsheet to re-check.
Built for Polish accounting firms and in-house finance teams
- KSeF XML parsed natively, alongside PDFs, scans and international formats
- Bank statements from ING, mBank, PKO BP, Santander, CAMT.052/053, MT940
- Four-dimension scoring — invoice no. 45%, vendor 30%, amount 20%, date 5%
- Split payments, fuzzy vendor names and amount tolerances handled by the engine
- Duplicate invoices and amount outliers flagged before they reach the books
What it is
InvoiceMatch is invoice reconciliation software for accountants who currently spend the first week of the month comparing a bank statement to a folder of invoices. You upload both sides in whatever shape they arrived — structured KSeF XML, a PDF, a photo of a receipt, a statement export from ING or mBank — and the engine pairs them up.
What makes it usable is that it argues its case. Every suggested match carries a score, and every score opens up: the invoice number contributed 45%, the vendor name 30%, the amount 20%, the date and currency the last 5%. One payment covering four invoices, a vendor spelled three different ways, a transfer that landed 40 groszy short — those are the cases the engine is built for, and the ones it explains rather than silently guessing at.
Files are held for 30 minutes and then deleted. Only anonymised counts stay in the log.
How it works
- Upload both sides. The month’s invoices — KSeF XML, PDFs, scans or photos — and the bank statement exactly as the bank exported it.
- The engine scores every pairing. Each candidate match is weighted across invoice number (45%), vendor (30%), amount (20%) and date and currency (5%). Split payments, vendor names written three different ways and small amount differences are resolved inside the scoring instead of being left for you.
- You review the exceptions. Confident matches are ready to confirm. What remains is a short queue, shown side by side with both documents and the reasoning behind each score.
- Export and move on. Audit-ready Excel exports go to the books; the processing history keeps match-rate trends over time.
Formats and bank integrations
On the invoice side InvoiceMatch parses KSeF XML natively and reads PDFs, scanned images, photographed receipts, international invoice formats and legacy exports. On the bank side it accepts statement exports from ING, mBank, PKO BP and Santander together with the ISO 20022 CAMT.052/053 and SWIFT MT940 standards — more than seven statement formats in total. Nothing has to be re-typed or converted before it goes in.
Security and data
Uploaded files live for 30 minutes and are then deleted; after processing, only anonymised counts remain in the log. Every generated result carries the transparency disclosure required by Article 50 of the EU AI Act. The design assumes the data belongs to an accountant’s clients — GDPR by design, not as an afterthought.
Who uses it
- Polish accounting offices (biura rachunkowe) running books for dozens of clients at once
- In-house finance teams of 1–10 people who own the AR reconciliation
- Founders still closing their own month, usually late at night
The common thread: people whose month-end is limited by typing speed rather than judgment.
What’s in the box
- Document extraction across KSeF XML, PDFs, images and legacy exports
- Seven-plus bank statement formats, including CAMT.052/053 and MT940
- Four-dimension match scoring with an expandable breakdown on every row
- Anomaly detection for duplicate invoices, amount outliers and suspicious patterns
- Exception queue with side-by-side document review
- Audit-ready Excel exports and a processing history with match-rate trends
- EU AI Act Art. 50 disclosure on every generated result
- GDPR by design — 30-minute file TTL, deletion after processing, anonymised logs
Frequently asked questions
Does InvoiceMatch read KSeF invoices?
Yes. KSeF XML is parsed natively, alongside PDFs, scans, photos and international formats — a single month can mix all of them.
Which bank statement formats are supported?
Exports from ING, mBank, PKO BP and Santander, plus CAMT.052/053 and MT940 — more than seven formats. The statement goes in as the bank produced it.
How is a match scored?
Across four weighted dimensions: invoice number 45%, vendor 30%, amount 20%, date and currency 5%. Every row expands to show how each dimension contributed, so a match can be checked rather than trusted.
What happens to my files?
They are deleted 30 minutes after upload. Only anonymised counts remain in the processing log.
Is InvoiceMatch available in Poland?
Yes. On the Polish market the same product ships as RozliczTo, at rozliczto.app.