Traveling with a chatbot: How to adapt chatbot personality to different traveler types

Abstract

Train travel is a popular way of transport in the Netherlands. Multiple tools exist for travelers to help them plan their journey and to look up information about their trip. However, these tools do not account much for the personality and behavior of the traveler, which might have a substantial impact on their preferences and needs. This thesis will focus on the different preferences and needs of different types of travelers when providing input through a chatbot meant for Dutch Railways, such as questions and complaints. The chatbot is meant to adapt based on the type of traveler using it. A Wizard of Oz experiment was conducted to see if traveler type has an impact on the preferences towards chatbot behavior. Some patterns could be observed, especially among the most outspoken traveler types, but further research with more participants is necessary to draw any decisive conclusions.