Investigating the Impact of Prosocial Compensation in Service Robot Failures

Mir, Mahmood Mustahsan (2025) Investigating the Impact of Prosocial Compensation in Service Robot Failures. Doctoral thesis (PhD), Manchester Metropolitan University.

Preview

Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (18MB) | Preview

Abstract

Despite the hype surrounding the operational benefits of service robots, their susceptibility to service failures and limitations in recovery remain critically underexplored. Existing humancentric recovery frameworks fall short in addressing service failures of emotionally agnostic robots, and cause significant customer dissatisfaction, financial losses, and abandonment of service robots in tourism and hospitality industry. To address this critical issue, this thesis investigates prosocial compensation as an alternative approach to service recovery in robot induced failure scenarios. Contrary to the dominant discount-based compensation approach applied during service recovery, robotic prosocial compensation offers an altruistic method of compensation that aims to safeguard customer satisfaction, business profitability, and societal wellbeing. To assess the efficacy of robotic prosocial compensation, three scenario-based, between-subjects experiments were conducted using hypothetical service failure vignettes administered online via Prolific. Study 1 employed a quick-service restaurant vignette and tested a single-factor design contrasting prosocial compensation (PC) with a monetary discount. The data were analysed using one-way ANOVA to assess the main effect of compensation type on post-recovery satisfaction. To enhance the generalisability of findings, Study 2 replicated the investigation in a hotel service context and introduced failure severity as a moderator. This study adopted a 2 (compensation type: (prosocial compensation vs. discount) × 2 (failure severity: low vs. high), between-subjects experimental design. The resulting data was analysed using a two-way multivariate analysis of variance (MANOVA) to test for both main and interaction effects across multiple outcome variables (satisfaction, empathy, moved, tolerance). In Study 3, a 2 × 2 between-subjects experimental design was also employed to systematically manipulate the type of prosocial compensation (high vs. low) and the type of robot voice (humanlike vs. robotic). The data were analysed using a moderated mediation approach, specifically employing Hayes’ PROCESS macro (Model 8) with 5,000 bootstrapped resamples, to test whether the effect of compensation type on customer satisfaction was mediated by perceived empathy, moved, and tolerance, and moderated by voice type. Across the three empirical experiments, this results consistently demonstrate that an emotionally resonant service recovery approach in the form of PC can significantly enhance customer satisfaction compared to the conventional transactional method of monetary discounts. Moreover, the study challenges the assumption that robots cannot be perceived as empathetic agents, by demonstrating that the integration of PC into recovery efforts leads to a shift in customer perceptions, making robots appear more emotionally engaged. Furthermore, this thesis explores the role of anthropomorphic features, specifically humanlike voices, in enhancing the effectiveness of PC. Robots with humanlike voices foster greater customer satisfaction, are perceived as more empathetic, elicit stronger feeling of being moved, and enhance customer tolerance towards service robot failures. Collectively, the thesis advances theoretical frameworks on service recovery within human-robot interactions and offers actionable insights for organizations aiming to enhance customer satisfaction.