This doctoral research explores the lack of user engagement observed in human-AI chatbot interaction through the theoretical lens of service-dominant logic (SDL). A recent observation suggests that stand-alone voice assistants and chatbots enabled by artificial intelligence are experiencing a fast-growing market adoption, especially among the user group of the Millennials, but the data published by Accenture in 2017 reveals that their use is limited to very simple tasks, such as turning on music or checking weather forecast only, and a frustrating user experience with these devices has been causing an ongoing challenge of user engagement. Although there is a great number of previous studies that examined the phenomenon of lack of user engagement of wearable devices, the studies that inquired into the phenomenon of lack of user engagement of other smart service products (SSPs), such as voice assistants and chatbots, are still limited. Also, most current studies focused on user adoption and acceptance of SSPs (Bolen, 2020; Deghani et al, 2018); therefore, understanding of the post-adoption user behaviour remains limited.

In order to address this knowledge gap, this doctoral research takes a research through design approach to inquire into the phenomenon of lack of user engagement in human-AI chatbot interaction. This research is an interdisciplinary study that is situated at the intersection of design and services marketing. Many recent contributions in the human-computer interaction (HCI) discipline that examined the challenge of user engagement generated new knowledge on how sensory stimuli can be leveraged to tackle the challenge of user engagement in a smart service system. However, the focus of these studies is significantly skewed towards visual, auditory, and most recently, haptic senses. Extant literature in sensory research and most recently HCI (e.g. Obrist et al, 2017; Spence et al, 2017) addressed the lack of studies that explore the HCI phenomenon through the olfactory sense, although the olfactory sense has the most intimate linkage with the areas of the human brains that directly control human emotions. Recent studies have started to respond to this knowledge gap, and the olfactory sense is gaining research traction, but most of these studies adopted a quantitative research approach that quantifies human experience as a measurable unit. Also, they do not offer deep insights as to how a designer can use the olfactory stimulus as a design material to design for an olfactory-enabled aesthetics of experience in the context of human-SSP interaction.

In order to expand the current knowledge in the effect of the olfactory stimulus on human emotions and experience, specifically in the context of human-AI chatbot interaction, this doctoral research explores the research problem - the lack of user engagement that leads to user abandonment in human-AI chatbot interaction. Based on the review of extant literature, this research problem was further developed into research questions as follows: 1) How could service designers use the olfactory stimulus as a design material to design for the olfactory-enabled aesthetics of experience in human-AI chatbot interaction? 2) Could the olfactory-enabled aesthetics of experience emerged in human-AI chatbot interaction enable value co-creation between the user and AI chatbot that enhances user engagement?

The contributions from this research are three-fold. First, this doctoral research is an interdisciplinary study that directly responds to the call for further research contributions from interdisciplinary studies by the service research community (Gustafsson and Bowen, 2017). Secondly, this research studied the olfactory stimulus as a design material that has been underexplored in the design and HCI disciplines. Finally, this research examined the lack of user engagement and user retention problem in a smart service system through the theoretical lens of SDL, whilst it adopted a research through design approach to explore rich details of human emotions and behaviour in the context of human-AI chatbot interaction. This study has examined the challenge of user experience qualitatively using a research through design approach, and this hybrid approach in combining the theoretical foundation from the services marketing discipline and a methodological approach from the design research discipline is original.

As a result, it expands the discipline-specific boundary of a design experiment by positioning a design experiment as a legitimate research method for exploratory research to scholars from non-design disciplines. This approach also demonstrates how service design can be used as a mid-range theory to operationalise meta-level theory - SDL - and explores an empirical observation - the lack of user engagement in human-AI interaction that is observed at a microlevel of a smart service system.