K+F Application support

 

Beneficiary: Nitro Communications Limited Liability Company

Project title: Development of an artificial intelligence-based customer segmentation system

Project ID: 2019-1.1.1-PIACI-KFI-2019-00413

 

Project description

 

Nitro Communication Kft, as a member of the Lead Generation group of companies specialised in the field of data-driven marketing and sales support, and Cessio Zrt., a consortium of receivables management companies, are developing a business intelligence solution based on the data assets of corporate customer communication and data repositories, exploiting the opportunities offered by artificial intelligence. The technology resulting from the consortium's R&D will be closely aligned with Nitro Communication's strategic plans in all of its target areas, while Cessio Zrt. will be able to directly benefit from the project's results in its business. Accordingly, a technology development process and a concrete product development process will be carried out in parallel within the framework of the project.

Any machine learning system is based on appropriate databases, on which the development of a targeted solution can be implemented. In this project, this basic structure is the data generated in the systems of Cessio Zrt., the most valuable part of which is the recorded communication with customers. The product resulting from the project aims to build on this extended understanding of the data assets to create a segmentation model that can help address two of the most important issues related to the debt management customers: solvency and, most importantly, willingness to pay, which are two of the most important conditions for debt settlement. Accordingly, our first task in this research is to compile an AI-interpretable database, with sources in the form of case management systems, recorded telephone conversations and other digital communication content used in customer contacts. The proposed solution is designed to process the data assets using semantic analysis tools, for which the content generated by all communication channels will be converted into a uniform format. Another important source of information for the system is the metadata beyond the content elements of the communication. The research will develop a metadata generation methodology to evaluate the metacommunication elements of the communication. In addition to recorded conversations, messaging services, but also metadata generated during web browsing, offer a realistic opportunity to extract metacommunication data. The resulting databases form the basis of our Machine Learning solution. Our goal is to create an AI engine that can segment customers by running on this data. The purpose of segmentation is to give the debt manager an indication of which customers are more likely to be settled and which customers are less likely to be settled. This information is of particular importance in the industry because companies such as Cessio Ltd. operate worldwide by purchasing bad debt portfolios from various banks and service providers for a fraction of the amount of the claim and only become profitable if they can achieve debt settlement for more than this amount. In a resource-constrained environment, he therefore stressed the importance of being able to select from a large customer base those likely to recover a debt.

Given that the proposed system is essentially based on a psychological profiling of the target function, it is expected that by modifying the target function, the same methodology can be successfully applied in other areas, such as sales or marketing.

Start of project implementation: 01.01.2020.

Physical completion of the project: 31.12.2022.

Aid amount: HUF 236 577 360