RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

AlfaStrakhovanie and Smart Engines have created a domestic AI system for classifying documents

Customers: AlfaInsurance SG

Moscow; Insurance

Product: Smart Document Engine (ранее Smart DocumentReader)

Project date: 2023/12  - 2024/06

2024: Building a system to classify and route documents

AlfaStrakhovanie and Smart Engines have created the first fully domestic artificial intelligence-based system designed to classify and route documents. It allows you to recognize, type and sort over 20 thousand documents per day. The solution was created as part of the import substitution program for foreign software. And speeds up the sorting of documents by 25%. Smart Engines reported this on July 31, 2024.

The system consists of two key components - proprietary OCR technology for full-text recognition of Smart Engines scientists and compact linguistic models created by AlfaStrakhovanie specialists. The product is 100% import-independent, since it does not use Open Source code or any other foreign software components.

OCR Smart Engines is capable of processing up to 15 pages per second. This speed is achieved through the use of ultralight or 4.6-bit neural networks - this is Smart Engines' own patented development.

AI recognizes and classifies any A4 documents: rigid (contracts, financial documents, client requests and much more) and flexible forms, photos and scans. In addition to Russian, the software solution provides recognition of text data in more than 100 languages ​ ​ of the world: from those based on Cyrillic and Latin alphabets to more complex ones - Arabic, Korean, Japanese, Chinese, etc.

File:Aquote1.png
Every day, AlfaStrakhovanie receives tens of thousands of documents: contracts, financial documents, client requests and many others. They must not only be recognized, but also typed, sorted, sent to the desired departments or provided at the request of tax and other regulators. The key factor for the transition to Russian OCR technology from Smart Engines for us has become the speed and quality of recognition. As a result of the implementation, we managed to noticeably - by 25% - speed up the processing of documents and minimize the risk of errors, "said Natalya Koretskaya, director of the operational department of AlfaStrakhovanie.
File:Aquote2.png

File:Aquote1.png
After the departure of foreign vendors, our company in time trouble created and brought to the market a text recognition system. AlfaStrakhovanie's results demonstrate how a motivated and qualified customer team can solve complex problems using Russian-made solutions. Our joint solution not only made it possible to solve the problem of the departure of Western players, but also turned out to be very elegant in terms of implementation, "said Vladimir Arlazarov, CEO of Smart Engines, Doctor of Technical Sciences.
File:Aquote2.png