How Kros uses Rossum to slash document processing times by 93%
time saved per document
average manual time per doc
KROS a.s. is a Slovakian software development company offering solutions for accounting and invoicing. In its 25+ years on the market, the company has always aimed at innovation and the newest technology to optimize the customer experience and retain its expert status. In recent years, the team at Kros has noticed an increase in the digitization efforts amongst its customers. The modernized legislation also allowed the possibility of converting paper documents such as invoices, bills, orders, or delivery notes into digital forms. These changes made it easy to store huge volumes of documents running through the hands of Kros’ customers.
Kros observed that all accounting software users struggle with two main needs: minimizing the time spent processing high volumes of documents and quality of extraction. The human error that often occurs is the biggest challenge in this industry as the complexity and strictness of the Slovak accounting system offered no space for mistakes.
The biggest challenge for KROS was to create a single workspace for customers to quickly and securely cover multiple document processing stages. These stages typically include the processing within the accounting software, transferring the data between Kros software and their clients, and storing the processed documents in a compliant way.
KROS offers web- and desktop-based accounting and invoicing solutions to its clients.
Because of the sensitive nature of the accounting documents, the most critical processes are run by the on-premise solution, where the new document processing AI would be plugged in. While Kros was already offering the ability to store accounting documents securely within its software, finding the right solution for document transfer, data capture, and processing had proven to be more challenging.
KROS selected Rossum as its partner of choice for several reasons: the ability to process documents in multiple languages, user-friendly validation interface, and smoothly integrate sensitive data into Kros’ own software in a compliant way and according to all legal regulations.
Rossum was integrated into Omega, one of Kros’ accounting software products. It is an on-premise solution designed to serve small and middle-sized companies. During the implementation period, KROS enjoyed the support of Rossum´s solutions engineering team, which helped overcome initial challenges and technical issues. Due to the sensitive nature of most financial documents, the integration was carefully fine-tuned to ensure compliance with the security needs of the end users.
Once the solution was in place, Rossum allocated R&D resources to support KROS’ efforts in explaining the processes behind the AI to the customers. This helped establish the KROS customers’ trust in the system in the early stages of deployment.
Since Rossum’s implementation, Kros observed more customers adopting the integrated solution. Many are happy to have a reliable solution that allows digital document processing at scale without needing to handle each supplier and business partner individually. To set up the end-to-end process for Omega users, the development and support teams of Omega are now using Rossum exclusively.
While in the early stages of the partnership, Kros’ customers used the system primarily to automatically process incoming invoices. Later, users started to use the system to process outgoing documents. Combined, Kros can now save its clients 93% of the time previously needed to process high volumes of documents and post invoices to the relevant downstream systems.
In the first half of 2022, Kros had seen the volumes of documents processed grow 5x, from 100.000 to 500.000. To offer its client greater flexibility, KROS is currently transferring its accounting solutions from on-premise to exclusively cloud-based. The next step is incorporating Rossum into each Kros cloud solution to offer an even more accessible and on-demand workflow.
“AI systems are learning, and they are not always perfect right out of the box. However, with little time invested into training, the number of errors is much lower. The initial investment into a reliable AI system is well worth the result!”Eva Mičurová, Software analyst at Kros