Structured document reader

Natural language processing and computer vision practices enable automated extraction of structured data from invoices or other structured documents such as test results or forms. Our solution saves the results in databases, eliminates errors, and utilises qualified personnel on tasks with higher added value.

Business case

Processing the documentation takes plenty of time, but the procedure could be automated, accelerated, and significantly improved by using innovative practices.

Our solution

Invoice data capture is naturally connected to several business functions. The receipt of an invoice triggers a series of processes with specific data requirements. Several data fields must be localised, and data extracted from those fields. The data capture process uses advanced machine learning to accelerate all steps and ensure accurate results.

End-to-end semi-supervised invoice reading solution based on experience with computer vision, Named-entity recognition, and natural language processing underlie the solution.


  • Machine learning-based document reader saves time and costs, reduces mistakes and enables a higher level of control over your data.
  • Increased data integrity give you a better insight into the process and makes it easier to keep track.

Need to Know More?

Ask us anything

Key contacts

Jakub Štěch - DataSentics

Jakub Štěch

Innovate Tribe Lead | Data scientist

Lukáš Tryner - DataSentics

Lukáš Tryner

Computer Vision Architect