OCR document processing & alerting

The logistics service provider was processing 800+ incoming documents a day by hand — delivery notes, invoices, consignment notes. Classifying and forwarding them took up three full-time positions every day.

RALUS trained an OCR and classification model on the existing document archive. Incoming documents are classified automatically, relevant data is extracted and fed into the target systems.

For exceptions or a low confidence score, a Teams alert goes out immediately. Manual rework is limited to genuine edge cases.