O2 Slovakia: Efficient analysis of customer’s chat communication

For many companies, instant contact with customers via the web or mobile applications is the key to improving their services. Chat communication is becoming increasingly popular. Quality evaluation and reporting on a vast number of conversations that customer support staff performs requires an automated solution. The nature of chats is a relatively big challenge from the natural language processing (NLP) point of view. 

In cooperation with O2 Slovakia, we prepared a solution that automates classification of chats that O2 customers have with customer support center staff. We researched and developed methods that deal with the “rough” language, typos, presence of domain-specific terms, and usually missing context. We evaluated statistical approaches as well as machine learning methods. We also used the latest advances from the so-called deep language learning and modeling.

Our common goal with O2 Slovakia goal is to better understand customer needs and ultimately improve the services provided via the gained knowledge and developed methods.