Automating Customer Experience Agents’ Evaluation with Natural Language Processing

Abstract

The importance and relevance of online commerce in society have grown uninterruptedly over the last years, with people changing their purchasing habits. The COVID-19 pandemic significantly accelerated this adoption, as quarantines forced people to stay in their homes. Under this scenario, online shopping became a more convenient option, considering product availability, the possibility to compare prices and shipping options, and door-to-door delivery. With this, e-commerce retailers and marketplaces must give excellent customer service to their customers, especially in this context when most activities are returning to their pre-COVID levels and restrictions. To manage customers that have any kind of problem, digital marketplaces and retailers usually have a team of Customer Experience (CX) agents that interact, talk and answer clients’ questions and inquiries, avoiding any potential negative situations on three different channels: voice calls, e-mails, and chat messages.