An Automated System for Identification of Tweets Requiring Customer Service Concern
Manoj Kumar Sethi, Kunal Batra, Manali Biswas, Mehar Lamba · 2022 3rd International Conference for Emerging Technology (INCET) · 2022
Social Network Communication has become an important factor for businesses for their successful operation. It has become important to connect with customers and offer support to them on these platforms. Several customers prefer to post comments, suggestions or complaints about a company’s products and services to online media such as Facebook, Twitter etc. because it is convenient for them and it increases pressure on product owners/service providers to respond. Twitter is a prominent social media platform with 206 million daily active users. Research from Twitter and Sprinklr suggest that 2 out of 3 people prefer Twitter for customer service over other social media platforms. It pays off for businesses when they attend to customer concerns or social media since studies show that people who receive customer service on Twitter from telecom brands preferred those brands 3 times more than those that had not received support. Yet many businesses aren’t connecting with consumers on Twitter. In this paper two unsupervised classification models are proposed that performs clustering on customer support dataset on twitter in such a way that companies can segregate those messages/tweets in which customers are needing genuine help. This will help the companies to actively respond to these customers and this can help increase their customer satisfaction.