Business challenge
Finance and accounting departments were losing hundreds of working hours every month manually reviewing, retyping and validating incoming invoices received through different channels (paper, PDF, email) before booking them into the ERP system, with frequent data entry errors.
DATUM solution
Using advanced cognitive OCR and machine learning, the DATUM Agent independently "reads" unstructured documents in any format. The system precisely extracts key entities (supplier, IBAN, tax amounts, dates, individual line items) and performs automatic 3-way matching. It prepares a structured dataset that is automatically injected into the accounting system for final approval.
Business value
Processing time per invoice was cut from an average of 8 minutes to under 1 minute. Human data entry errors were almost completely eliminated and the monthly financial close was accelerated.