Rossum is an AI document processing platform for transactional paperwork such as invoices, purchase orders and delivery documents. AI agents read each document, capture and validate its data, send emails, ask for approvals and write the result to an ERP. A proprietary transactional model handles 276 languages and handwriting, and a 14-day trial and free demo are offered. Rossum is now part of Coupa.
Rossum is built for finance and operations teams drowning in inbound documents. Accounts payable groups are the main audience, followed by supply chain management and order management teams. Enterprises with high invoice volumes, shared service centers and multi-country entities get the most from it, as do systems integrators who connect it to an ERP. Business users review exceptions in a unified interface, while administrators set rules, approvals and data policies. Procurement and IT teams evaluating document automation can start with the trial and a guided tour, then involve systems integrators for ERP connections.
Typical workflows start with documents arriving by email, scanner, PEPPOL or a shared drive. Rossum extracts the fields, cross-checks them against master data and business rules, and infers values such as general ledger, tax or customs codes. Clean transactions pass straight through, and exceptions trigger automated vendor notifications. Customers report concrete gains: Wolt cut error rates by 44 percent, Adyen reached 92.6 percent accuracy after only 20 documents, and Fugro reduced invoice processing time by 70 percent. Dashboards track turnaround time and straight-through processing rate, and every document keeps an audit trail.
