Improving text legibility of a chatbot conversation to gain customer trust A quantitative study on typeface curvature and the effect on trust

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
Chatbots in the marketing industry are increasingly acquiring businesses' and brands' attention to connect and learn directly from their customers. However, chatbots' best practices to exploit their full potential are little known. The absence of knowledge on chatbot best practices has led companies to study diverse marketing strategies to increase user acceptance of the tool. For example, studying the correct use of aesthetic elements of a conversation to enhance the customer experience. This study examines if the use of curved typefaces during a chatbot conversation is perceived as more legible by the participants than angular typefaces and to what degree it increases the user’s trust in the chatbot in sequence. The research included participants above 18 years old who interacted with a fictive bank chatbot that communicated either with a curved typeface or an angular typeface. The participants evaluated the conversations via a survey. The results indicated strong support for a mediation model. Meaning that the use of curved typefaces in a chatbot conversation positively affects customer trust, and `legibility` positively mediates this relationship. The perceived risk of the service type context was researched as a moderator of the analysis to determine whether low or high levels of perceived risk affect the relationship between the variables. This was achieved by creating two different contexts of the conversations between the participants and the bank chatbot. A high-risk scenario was created where participants had to share personal information with the chatbot; the low-risk scenario did not involve risky information sharing. Nonetheless, results showed no support for the moderated mediation, meaning that perceived risk did not moderate any of the relations. This research implies that curved typefaces during a chatbot conversation are perceived as more legible and, in sequence, increase the user's trust in the chatbot.
Faculteit der Managementwetenschappen