TIP: You can scan multiple invoices into one PDF and Datamolino will split the invoices for you. Also, you can ask suppliers to email you PDF invoices which you can forward to your Datamolino account.

Each business is unique, that is why Datamolino first learns to read your files during a "fingerprinting process". 

When we receive a document, we match it against our database - the system checks if it can match a fingerprint. 

  • If it is a new layout of a document, we set it up in our system - create a new "fingerprint".
  • If there is a "fingerprint match", the document is processed automatically without human interaction in Datamolino. This means that the data extraction takes only moments.

Datamolino works best when we had a chance to learn the invoices that you typically process. 

The system works best with electronically generated PDFs (ask your suppliers to send you a PDF instead of paper) or when the scanned documents have good optical quality (set your office scanner to 300dpi). With scanned documents (PDF, JPEG, TIFF) we are limited by the capabilities of OCR. There may be occasions, where a human may be able to discern the values on the receipt, but the OCR fails to pick up anything. This happens mostly on pictures taken with smartphones. The quality of the image is not sufficient.

After Datamolino learns to extract data from your invoices, the data extraction takes only moments. The best way to test Datamolino is to upload older invoices, wait until they are setup and then upload newer invoices from the same suppliers and see how Datamolino performed for you.

WHAT CANNOT BE AUTOMATED (with these types of documents we try our best manually, which means you have to wait for the results)

  • paper was folded in multiple places before scan and the folds interfere with OCR
  • paper is not a scan but a photo, or low DPI scan
  • handwriting interferes with OCR
  • low resolution
  • document is hand-written
  • the card transaction receipt overlays the invoice


#fingerprinting #scan #quality

Did this answer your question?