Who is more forgivable? The interaction effect of consumption type and service failure type on guest forgiveness

Purpose In the tourism and hospitality industries, contemporary guests frequently encounter two types of service failures: those caused by artificial intelligence (AI) and those caused by humans. This study aims to explore the interaction effect of consumption type (hedonic [pleasure-driven] versus utilitarian [function-driven]) and service failure type (human versus AI) on guest forgiveness based on the expectancy disconfirmation theory. Design/methodology/approach A total of 696 participants completed the three experiments to test the proposed hypotheses. Findings The findings revealed a significant interaction effect: when guests engage in hedonic consumption rather than utilitarian consumption, human service failures, as opposed to AI failures, result in lower forgiveness. This interaction is mediated by disappointment. Furthermore, this research demonstrates that this interaction is moderated by regulatory focus, being stronger for prevention-focused individuals. Practical implications These findings offer valuable insights for tourism and hospitality enterprises using both AI and human service agents. Originality/value These findings provide a novel understanding of whether AI or human service failures result in more negative outcomes, which determines how managers subsequently adopt appropriate coping strategies.