The dark side of AI-powered service interactions: exploring the process of co-destruction from the customer perspective

Abstract

Artificial intelligence (AI)-powered chatbots are changing the nature of service interfaces from being human-driven to technology-dominant. As a result, customers are expected to resolve issues themselves before reaching out to customer service representatives, ultimately becoming a central element of service production as co-creators of value. However, AI-powered interactions can also fail, potentially leading to anger, confusion, and customer dissatisfaction. We draw on the value co-creation literature to investigate the process of co-destruction in AIpowered service interactions. We adopt an exploratory approach based on in-depth interviews with 27 customers who have interacted with AI-powered chatbots in customer service settings. We find five antecedents of failed interactions between customers and chatbots: authenticity issues, cognition challenges, affective issues, functionality issues, and integration conflicts. We observe that although customers do accept part of the responsibility for co-destruction, they largely attribute the problems they experience to resource misintegration by service providers. Our findings contribute a better understanding of value co-destruction in AI-powered service settings and provide a richer conceptualization of the link between customer resource loss, attributions of resource loss, and subsequent customer coping strategies. Our findings also offer service managers insights into how to avoid and mitigate value co-destruction in AI service settings.

Hybrid Service Recovery: Design for Seamless Inquiry Handovers between Conversational Agents and Human Service Agents

“…We contribute prescriptive knowledge about form and function with an initial set of design principles for research to guide the interlocking of CAs and HSAs to improve service recoveries, which incorporate request escalation [13,19]. We also contribute to service literature by addressing the identified need in literature for improved service recovery strategies for CAs involving HSAs [12,13,17,32,38]. Established fallbacks involving request escalation are extended by providing a holistic service recovery strategy for CAs, addressing the challenge to provide real-time support and incorporating valuable insights from theory and experts in the field of customer service.…”

“…To avoid negative effects in terms of customer dissatisfaction due to additional waiting times, these solutions should address the challenge of realtime support [19]. The seamless handover of requests from CAs to HSAs as a recovery strategy addresses the suggestion in the literature of transferring requests to employees, if CAs' capacities are exceeded [12,17,32,38].…”

“…To fully take advantage of CAs and ensure high quality service, scholars call to investigate conditions, implications of and recovery strategies for service failure [15][16][17]. This knowledge is especially important for the prevention of detrimental effects connected to possible failures by CAs (e.g.…”

“…Thus, scholars call for service recovery strategies to compensate technological boundaries during service delivery and avoid customer dissatisfaction [15][16][17]. In this matter, previous research has focused on CA-initiated repair strategies to avoid conversational breakdown, ensure continuation of the dialog process and successful outcome [18,19].…”

“…To realize this hybrid service recovery strategy, the design has to consider the socio-technical interplay of technology, processes and humans (employees and customers) and goes beyond technical system specifications [21][22][23]. To prevent customer frustration and support HSAs, relevant information of the CA-customer-interaction should be provided for the handover so that service process steps and questions do not have to be repeated [13,17]. Therefore, connected processes are required to promote seamless handovers from CAs to HSAs [24,25].…”