When robots fail: Dual pathways of employee appraisals in hospitality

AI-enabled service technologies are increasingly deployed in hospitality and other customer-facing industries, yet failures of these systems create not only customer dissatisfaction but also significant challenges for employees. This research develops and tests a dual-pathway framework to explain how employees interpret and respond to AI service failures. Drawing on cognitive appraisal theory and the challenge–hindrance framework, we propose that failures are simultaneously appraised as challenges, which foster problem-solving, and as hindrances, which foster cynicism. Across three complementary studies, we provide cumulative evidence for these mechanisms. Study 1 employed a laboratory experiment with hospitality employees (N = 120), who viewed video vignettes of a humanoid robot performing successfully or failing, after which they completed validated scales of challenge and hindrance appraisals. Study 2 used an online scenario experiment with hospitality employees (N = 200), where participants read service failure or success scenarios and responded to multi-item measures of appraisals, problem-solving, and cynicism. Study 3 surveyed frontline hotel employees (N = 262) in East Asia, collecting self-reported data on the frequency and severity of AI service failures and employees' appraisals and behaviors, along with validated measures of technology self-efficacy and job autonomy. Taken together, the findings demonstrate that AI service failures simultaneously stimulate constructive engagement and destructive withdrawal, depending on employees’ appraisals and resources. This study extends service failure research beyond its customer-centric tradition and highlights actionable levers for managing employee responses in AI-mediated service ecosystems.