Rossum

Rossum is a leader in intelligent document processing, offering an AI-powered platform that automates the extraction and processing of data from various business documents. Founded in 2017 in Prague, Czech Republic, by Tomas Gogar, Petr Baudis, and Tomas Tunys, Rossum's technology focuses on streamlining the handling of transactional documents such as invoices, purchase orders, and shipping manifests. The company's solution combines advanced data extraction capabilities with a low-code, cloud-native platform to significantly reduce manual work in document processing workflows.

At the core of Rossum's offering is its proprietary AI engine, which includes a specialized Transactional Large Language Model (T-LLM) designed specifically for business documents. This model, trained on millions of annotated documents, enables Rossum to achieve high accuracy in data extraction without the need for extensive training. In February 2024, Rossum unveiled Aurora, its next-generation AI engine, which further enhances the platform's capabilities, offering improved accuracy and faster processing times.

Rossum's technology typically saves users over 90% of manual work associated with document processing. The platform is designed to be language-agnostic, allowing it to operate globally and handle documents in multiple languages. This has made Rossum particularly attractive to multinational corporations seeking a unified solution for their document processing needs.

Key customers and partnerships

Rossum serves over 150 global clients, including prominent companies such as Siemens, Bosch, Veolia, EY, Adyen, Cushman & Wakefield, BDO, Suez, and PwC. These organizations utilize Rossum's platform to process a wide range of documents, from invoices to financial statements and customs declarations.

Adyen, a global financial technology platform, has implemented Rossum's solution to process over 1 million invoices and Boletos annually. With the introduction of Aurora, Adyen experienced a 20.5% reduction in errors and improved extraction accuracy to 93.4% on average.

Wolt, an express food and grocery delivery platform operating in 25 countries, uses Rossum to handle over 100,000 invoices yearly from more than 2,000 suppliers in 20 languages. After implementing Aurora, Wolt saw a 44% reduction in average error rates and an increase in straight-through processing of invoices from 46.5% to 58% within just two days.