The User Experience of Chatbots: A Design Science Approach

Zeljko Maric

Student thesis: Master thesis

Abstract

Even though Chatbots have been a research subject since the early 1960s when Joseph Weizenbaum developed the first text-based bot named ELIZA there is little research on the user experience of chatbots. Possible reasons are the overemphasis of HCI scholars on graphical user interfaces (Følstad & Brandtzaeg, 2017) and the fact that most of the mainstream messaging platforms like Facebook Messenger opened their platforms for developers just in the last two years. The question what a good chatbot user experience entails remains mostly unanswered. This thesis wants to make a first contribution in answering this question by following a Design-Science Research (DSR) approach. A chatbot prototype was build and then user tested with 10 students (age 22-29) through formative usability testing to explore factors that influence a good user experience. Additionally debriefing interviews were led to explore unvoiced aspects during the user test. Based on the insights from both user testing and debriefing interviews, user experience guidelines were formulated. First results indicate the users’ overall preference for human-like characteristics in a bot, but users also reported that some specific bot-like characteristics made the conversation preferable to a human-to-human interaction. Among other things participants highlighted the non-judgemental space provided by a bot that made them feel safe to voice otherwise unvoiced thoughts.

EducationsMSc in Business Administration and Information Systems, (Graduate Programme) Final Thesis
LanguageEnglish
Publication date2018
Number of pages90
SupervisorsTorkil Clemmensen