Artificial Intelligence in Services – Determinants of an Ideal Human-Technology Mix in Service Encounters
Artificial Intelligence (AI) and AI-based service robots are increasingly adopted in service encounters (Teixeira et al., 2017) which have significant impact on customer experience during service provision (Larivière et al., 2017; Ostrom et al., 2015). Although replacing human employees with AI-powered machines has several advantages such as increased efficiency (Huang & Rust, 2018; Wirtz & Zeithaml, 2018), customers loose the opportunity to obtain human service (i.e., contact to employees) which has been related to positive customer outcomes (e.g., customer delight) (Collier et al., 2018) and is relevant for customers’ evaluation of a service encounter. (Collier et al., 2018; Gremler & Gwinner, 2008). Moreover, in several service settings human service is preferred over technology (Rafaeli et al., 2017) and artificial agents that resemble a human too closely could be perceived as creepy and cold (van Doorn et al., 2017).
Hence, service firms face trade-off challenges when replacing human employees with AI (Matzner et al., 2018), begging the question what factors determine an ideal human-technology mix in service encounters (Larivière et al., 2017). Drawing on Service Encounter Needs Theory (SENT) (Bradley, et al., 2010) and the Robot Acceptance Model (RAM) (Wirtz et al., 2018), we develop three research questions:
- What distinguishes the application of AI-based technologies in service encounters in the eyes of the customer compared to other technological innovations?
- Which psychological needs are fulfilled by AI-based service agents and what characterizes services which specifically require human service?
- Which customer individual (e.g., technology anxiety) and contextual (e.g., service type) factors determine an optimal human-technology mix in AI-enabled service interactions?
We aim to make a three-fold contribution to literature. First, we unfold customer attitudes towards AI-based service agents. Secondly, we develop a framework for an ideal human-technology mix and reveal which individual and contextual factors shape customers’ perception, introducing new and context-specific antecedents and outcomes for AI-based service interactions. Lastly, we expand theory about SENT and RAM. Researchers might use our results to develop new and adapt existing models on service interactions, practitioners to strategically balance the application of AI-based agents.