Publication date:
Publication from:
Ukrposhta is expanding its use of artificial intelligence from IT-Enterprise to process primary documents.
The companies continue to scale their joint AI project: in just the first three months of 2026, the intelligent agent processed over 26,000 documents, fully automating routine verification and data entry operations. As a result, the error rate decreased to 0.5%, and financial period closing has significantly accelerated. The accuracy of intelligent document recognition has reached 99.5%.
For Ukrposhta, this implementation has addressed the large volume of incoming primary documentation that previously required substantial manual effort. Processing time for such documents has been reduced by 60–80%. Within the team, the AI assistant has been named Marko—a symbolic reference to the postal stamp and also one of the most popular modern names in Ukraine.
Previously, the finance team manually processed hundreds of supplier documents, which created a significant workload during closing periods. Today, the system automatically reconciles data with contracts and specifications, while specialists focus primarily on documents with discrepancies. This provides not only time savings but also a higher level of control and process transparency,
said Dmytro Didyk, Director of the Procurement Department at Ukrposhta.
A key challenge of the project was not text recognition itself, but the diversity of document formats. Suppliers submit documents in hundreds of variations—with different structures, wording, and levels of detail. Even the same service may be described differently. That is why the AI agent goes beyond recognition—it interprets content, matches it against contracts, and forwards only inconsistent documents for additional review.
At the same time, Ukrposhta updated its internal processes to ensure proper interaction between primary documents, contracts, and the ERP system, eliminating data duplication.
The Ukrposhta project has demonstrated that Ukrainian AI solutions can already handle large volumes of documents in mission-critical business processes. For large companies, this primarily means reducing routine workload, minimizing errors, and accelerating operational decision-making,
commented Oleh Shcherbatenko.
This case demonstrates the transition from experimental AI usage to systematic application in daily financial operations, delivering measurable impact in speed, accuracy, and workload optimization.
At the same time, IT-Enterprise continues to develop its AI ecosystem. During the presentation of Version 2026, the company announced 81 new AI solutions, already used by leading Ukrainian businesses, including WOG, YASNO, Epicentr, and the Come Back Alive Foundation.
