Abstract

Today, the availability of similar IT products and services to customers has made the telecoms market highly competitive which has led to an issue of customer churn due to their inability of being able to detect and predict the customers that are going to remain.

This research aim is to provide organisations with several Artificial Intelligence techniques on how to determine customer satisfaction and deter them from leaving. It has been proven that it is less expensive to retain existing customers than to engage with new customers because customers normally are faced with competitors throwing adverts, emails, and exclusive discounts that would encourage them to patronize their services/products.

This research has also explored the strategies to retain customers by surveying the causes of customer dissatisfaction using reviews coming from customers which are considered one of the major approaches for customer satisfaction. The use of data mining/prediction tools such as Weka, and PowerBI were used for data visualization/ prediction, as well as using machine learning and deep learning models to determine various evaluation scores and how accurately they performed.

For the machine learning model, the Random Forest had the highest performance score with 80% while the Roberta model had an accuracy score of 92%.

To sum it up, an artefact was created to predict reviews so as to provide excellent guidance to organizations on how to improve customer satisfaction and deter churn rates based on the context of their words.

The Application of Artificial Intelligence techniques to determine customer satisfaction and churn rate - a case study of telecommunication