Understanding the user experience through interaction types in customer service chatbots

• Autor(a) principal: Barata, Pedro Miguel Santiago
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.5/30054
• topic: Chatbots; Virtual agents; Chatbots interaction; User experience chatbots; Customer service chatbots; Conversational chatbots; Rules-based chatbots; Chatbot performance; Chatbot usability; Chatbot tasks; Agentes virtuais; Interação de chatbots; Experiência de utilizador com chatbots; Chatbots de serviço ao cliente; Chatbots conversacionais; Chatbots baseados em regras; Desempenho de chatbots; Usabilidade de chatbots; Tarefas de chatbots; Domínio/Área Científica::Ciências Sociais::Outras Ciências Sociais

Understanding the user experience through interaction types in customer service chatbots. Chatbots have become increasingly popular in recent years due to their ability to automate tasks, improve customer engagement, and reduce operational costs for businesses. They are also seen as a way to improve the user experience by providing quick and easy access to information and services through messaging platforms that are already familiar to users. With the raising and appearance of this type of communication, this research aims to offer an insightful analysis of chatbot interaction types, their optimal applications, and their suitability for various tasks. It also seeks to provide guidance on key considerations and pathways for effectively implementing chatbot technology. In order to carry out this research study, we created four prototypes using Chatbot.com. In each scenario, we trained two chatbots: one using AI with text input and the other using rules with button input. Based on these methods, research questions were considered to further investigate the performance, usability, satisfaction and mental perception. To assess chatbot interaction types' performance in specific tasks, we use metrics like Customer satisfaction (CSAT), Net promoter score (NPS), Emotional response (SAM), Usability (SUS), or the Post-Task Workload (NASA-TLX). Overall, we find better results with the rules-based button input approach. We find and identify certain conditions that could lead into these results, such as task complexity or the trained approach. This study demonstrates that rules-based chatbots outperform other types in certain conditions and tasks. It underscores the importance of analyzing task-specific outcomes, mental workload, and complexity perception to ensure user satisfaction and protect brand reputation. Additionally, the underlying technology significantly influences chatbot performance and task completion.