THE IMPORTANCE OF NATURAL LANGUAGE RECYCLING IN AUTOMATING CUSTOMER SERVICE

Abstract

The use of artificial intelligence and natural language processing (NLP) in customer service is growing quickly. Technology is being used to interact with users and answer their questions. Using NLP in customer services is in the form of artificial intelligence applications that allow users to communicate with models using different languages through text or speech, and the model will provide answers to the users. The main goal of this systematic review is to locate and analyze the existing articles and studies on the use of NLP technology in customer service in terms of research domain, applications, datasets used, and evaluation methods. Also, this systematic review looks at the future directions in the field and where it is going in addition to the existence of any significant limitations. The time period covered by the study is from 2015 to 2022.  During the systematic review, all related papers were found, extracted, and analyzed using five major scientific databases. To create the final review article, relevant papers were sorted and filtered based on inclusion/exclusion standards and quality assessment. According to our findings, chatbots and question-answering systems were used in 10 main fields and mostly utilized in general, social networking and e-commerce areas. In addition, we discovered that Twitter dataset was the second dataset in terms of the most often used datasets. The majority of the research used their own original datasets in addition to Twitter dataset. For the evaluation, most of the researchers used Accuracy, Precision, Recall, and F1 as the methods to evaluate the performance. Also, future work is discussed including the need to improve the performance of the models and the size of the datasets used, as well as it aims to better understand users’ behavior and emotions is included. Moreover, there are limitations faced by the researchers and those limitations are diverse.