Rossum

Rossum is a document AI company founded in 2017 in Prague, Czech Republic, by a team of AI researchers who set out to remove manual data entry from paper and PDF business documents. It targets finance, procurement, and supply-chain teams that lose time and accuracy processing invoices, purchase orders, and customs paperwork by hand.

The platform combines a proprietary transactional large language model, said to support 276 languages and handwriting, with modules for document capture, data validation, automated communication with senders, and real-time analytics. Enterprise customers such as Wolt, Morton Salt, and the Port of Rotterdam use Rossum to automate accounts-payable and logistics document workflows, reporting reduced processing time and manual error rates.

1. Document capture: Extracts data from invoices, purchase orders, and delivery notes with a transactional LLM.

2. Data validation: Cross-checks extracted fields against business rules before they reach ERP systems.

3. Automated communication: Requests missing information from suppliers without human follow-up.

4. Multilingual support: Processes documents in 276 languages, including handwriting.

5. Analytics dashboard: Tracks straight-through-processing rate and time saved per document.

6. ERP integrations: Connects to the finance and procurement systems enterprise customers already use.