When Chatbots Fail: Exploring Customer Responsibility Attributions of Co-Created Service Failures: An Abstract

Abstract

Extant literature has predominantly viewed value co-creation as a voluntary process of value creation between the customer and the service provider (e.g., Grissemann and Stokburger-Sauer 2012; Sugathan, Ranjan, and Mulky 2017). However, the customer may not always enter into the value co-creation process voluntarily. As chatbots are increasingly replacing traditional customer service options (Belanche et al. 2020), customers may find themselves in situations where they are forced to use automated technologies (Reinders, Frambach, and Kleijnen 2015). Indeed, several service providers have started to offer chatbots as the only customer solution, especially as the first point of contact for customer service queries (Forrester 2017). Such cases, where the customer does not have a choice as to whether to interact with a chatbot or a human representative, represent instances of mandatory customer participation (Dong and Sivakumar 2017; Tsai, Wu, and Huang 2017).

Limiting consumers’ perceived freedom of choice in co-creating, may result in severe negative attitudes towards the evaluation of the technology and the service provider (Reinders, Dabholkar, and Frambach 2008). Such attitudes may be further exacerbated when a situation requiring the forced use of a technology leads to service failure. We argue that in cases of service failure, forced co-creation (no choice but to interact with chatbot) will have a distinct impact on customer responsibility attributions of controllability, stability and locus of causality, when compared to voluntary co-creation (choice between chatbot and human representative). We also investigate the role of customer expectations and propose that expectations mediate the relationship between each co-creation setting and the resulting responsibility attributions.

Experimental research is being proposed to test the research framework in a customer service setting, taking into consideration two types of service failure: outcome and process failure. Participants will be randomly assigned to two different co-creation settings: forced co-creation and co-creation by choice, and will be required to interact with an actual chatbot, which has been programmed using the IBM Watson platform. Participant responses will be recorded through an online questionnaire.

The findings of this experimental research are expected to contribute to a more insightful understanding of the consequences of forcing consumers to use novel technologies. This is an important research area which has been largely overlooked in service literature to date, as the literature has typically examined the introduction of new technologies within a voluntary context. Insights into these behaviours are also important to achieve a more substantive understanding of the implications of AI technologies, and to find ways to reduce negative reactions to such technologies.