Shaping Chatbot Efficiency—How to Build Better Systems?

Various types of AIArtificial Intelligence (AI) systems are distinguished based on the algorithms deployed, the technical features, and the devices integrated into different applications. The puzzling question hereby is whether these systems provide the desired experienceExperience and satisfactionSatisfaction to the userUser in regard to efficiency of chatbotsChatbot currently available on the market. As seen from the marketingMarketing examples and the profound literature audit reported in the previous chapter, chatbotChatbot efficiency perception and thus system adoption and use are very sensitive to userUser needs and demand for a satisfactory experienceExperience. As is well known from the behaviour theories, satisfactory experienceExperience fosters positive attitudesAttitudes and thus great willingness to use a product. From UXUser Experience (UX), we are also well informed that satisfactionSatisfaction is crucial for inspiring new computational and design frameworks for AIArtificial Intelligence (AI). Therefore, challenging fundamental assumptions on the factors driving attitudesAttitudes and satisfactionSatisfaction, we aim to provide the much-needed understanding of how to build better systems for AI chatbotAI chatbots implementation. In particular, qualityQuality and ease of useEase of use are discussed as core parameters loading on the way chatbotChatbot efficiency is evaluated. We further look at the factors shaping interactivityInteractivity. Both cognition and emotion turn to play a role. As functionalityFunctionality (cognitive) and enjoymentEnjoyment (emotional components) have emerged as the most frequently explored in various HCI, UXUser Experience (UX), and marketingMarketing studies, we focus on these parameters and their antecedents. While in previous research abovementioned issues have been addressed in separate studies, often in isolation, hereby we combine cognitive and affective components in a conceptual model that will be tested in empirical studies, as described in detail below.