Every year, the companies spend even several months to enter the data of accounting invoices. To make the process less time-consuming and reduce the risk of errors, our IT specialists, with the support of accounting professionals, developed a solution called AIGA – an artificial intelligence accounting assistant. Any kind of content that needs to be extracted from unstructured information in context is a problem for many companies, but we took it as a nice challenge that turned into a quality, needs-based solution.


We developed an automated invoice processing tool that can retrieve standard information from a document:

  • invoice number;
  • invoice date;
  • payment date;
  • supplier’s registration number;
  • supplier’s VAT registration number;
  • amount;
  • currency.

But that’s not all – the artificial intelligence algorithm can also be trained to retrieve and process other information. Solution AIGA can process an electronic image or PDF document, extracting all textual information from it, and analyze the obtained data to select the necessary values. To ensure high accuracy of the solution, the algorithm is trained with more than 50,000 different invoices.


AIGA solution not only allows you to significantly save time (up to 40 hours for entering 1000 invoices), but also to increase data security, as the data is not sent to the cloud, but is stored locally on the company’s servers. Also, the tool marks data in which algorithms determine low recognition accuracy, so that employees can additionally check and correct information if necessary – thus reducing the possibility of errors in accounting data.

The accounting assistant is adaptable to different company needs and serves as an excellent business model, as costs remain constant regardless of the volume of documents.