Do explanations in conversational agents in recommender systems affect consumer trust and purchase intention? : an experiment on chatbots

In order to meet the desire of online shoppers for personalized product recommendations, companies are increasingly using conversational recommender systems that encounter the customer in the form of chatbots, for example. Explanations have already proven their worth with earlier recommendation technologies such as expert or recommender systems to make the systems more transparent and increase trust. In conversational recommender systems, such explanations are usually based on product information or on information from online consumer reviews, which in turn results in two styles of explanation. Which of the two styles elicits greater trust among users and how this affects behavioral intentions in the field of e-commerce is as yet unexplored and is investigated in this paper with the help of an online experiment. 199 subjects participated in this experiment. However, the results of the applied regression analyses could not detect a significant difference between the two explanatory styles. Nevertheless, this study contributes to strengthening the understanding of explanations in conversational recommender systems in e-commerce through further hypothesis testing.
submitted by: David Burghold
Masterarbeit Universität Innsbruck 2023