Nowadays, bookkeeping deals with substantial amounts of paperwork that must be processed on a monthly basis. This two-way process requires both entering the data from documents into the systems and accounting for them accordingly, which is not only tedious and accountant-intensive but also prone to human error.
Datarabbit, a member of the PrediqTank consortium, is working with e-file, one of Poland’s top accounting software companies, to provide a solution that is error-free while requiring less effort and actual work from accountants.
Results & Client Benefits
DataRabbit's team uses the latest advances in image processing, such as neural networks, to implement an OCR system that automatically reads all text from invoices (electronic documents or photographs), and then extracts the information valuable from an accounting standpoint, greatly reducing the need for manual data entry.
As part of later stages, the pipeline can also automatically alert about anomalies/unusual situations that may require special treatment (such as unusual tax rates).
In addition to the development work, DataRabbit's team is responsible for deploying and integrating the solution with the client's Azure cloud environment.
Industry: Accounting
Project Category: AI/ML, image processing, software development
Lead Member: Datarabbit
Technology used: Python, Pytorch, Terraform, Azure