Decision-Making in Conversational Systems: User Response Time and Choice Overload on Chatbots

• Main Authors: Hsia, Po-Jen, 夏柏仁
• Other Authors: Ray, Soumya
• Format: Others
• Language: en_US
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/vq38fb

碩士 === 國立清華大學 === 服務科學研究所 === 106 === Conversational systems are highly growing recently with the increasing use of mobile devices, Internet of Things, and artificial intelligence. Conversational systems aim to communicate with humans in a dialogue flow style and imitate human-to-human interaction. This style of interaction makes conversational systems fundamentally different from traditional computer systems. In this study we are looking for how limitations of conversational interface impacts decision- making of users in chatbot context. We’ve designed an e-commerce chatbot for AsiaYo, a Taiwanese e-commerce firm that provides an accommodation platform for travelers. In collaboration with AsiaYo, we created a Facebook chatbot that invited potential customers to engage in a conversation to find an accommodation for an upcoming or future vacation in Asia. Given that e- commerce chatbots serve as an alternative to traditional menu-driven website and mobile applications, we investigated whether many of the traditional assumptions held true in a conversational setting. We wondered how the latency and limited display options of chatbots might affect users. The result shows that users are affected in interesting ways by latency of system and number of options displayed in chatbot. We then provided several theoretical and practical insights base on these relations to help improve future user experience design for chatbots.