RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

NLogic will create a system of automatic recognition of customer statements for Tele2

Customers: T2 RTK Holding (formerly Tele2 Russia AB, Tele2)

Contractors: NLogic
Product: Artificial intelligence (AI, Artificial intelligence, AI)

Project date: 2021/04  - 2021/09

2021: Development of a solution for automatic processing of applications from customers

The mobile operator Tele2 summed up the results of the open request for proposals for the development of a solution for the automatic processing of applications from customers. The developer algorithms machine learning company (NLogic part of the multidisciplinary IT group "") IKS Holding became the winner, receiving a contract for 5 years and announced this on November 9, 2021.

Large companies that are faced with a huge flow of documents and standard operations are growing the need for digital solutions and automation of routine processing processes.

Mobile operator Tele2 daily processes a large volume of applications and calls from customers. The company announced a tender for the development of a product to optimize such tasks.

The use of artificial intelligence technologies in recognizing the same type of applications will help reduce the time to make a decision on client requests and free employee time for more demanding tasks. So, the client will get the result faster, and the company will reduce the cost of servicing a labor-intensive process.

The main requirement for the system was the high quality of recognition of statements, including the ability to recognize files and the text contained in them in most formats; Support for automatic correction and rotation of images Work with files that contain multiple statements at the same time ability to recognize structural elements such as checkboxes, markers, etc. In addition, the system must meet the requirements for performance and information security, as well as provide the ability to easily integrate.

The company nlogic in just 1 month was able to prepare a project of the system using a team of three people and qualify for various criteria among six participants. At the same time, testing of the system was carried out on a deferred sample of 100 applications. Attribute recognition accuracy was 97%.