Blending minds and machines in service operations: enhancing managerial responses to online reviews through human-generative AI collaboration
Lin Jia et al.
Abstract
Purpose Generative artificial intelligence (GAI) is increasingly embedded in service operations, particularly in crafting managerial responses (MRs) to online negative and positive reviews. Yet, little is known about how firms should optimally allocate responsibilities between human staff and GAI across distinct response tasks, nor how such allocations shape consumer perceptions and downstream behavioral intentions. This study investigates how customers evaluate MRs produced by different agents including humans, GAI and multiple forms of human and GAI collaboration across specific response tasks. Design/methodology/approach We conducted a randomized online experiment. Main effects were examined using ANOVA with data from 879 participants, and the underlying mechanisms were further analyzed using the PROCESS macro with a parallel mediation model. Findings Our findings show that to improve customer satisfaction and increase booking intentions, negative reviews are most effectively addressed through an augmented human approach in which GAI creates an initial draft and experienced professional staff refine it. In contrast, positive reviews are best handled solely by experienced professional staff. Originality/value This study advances knowledge on how to combine GAI with human efforts in service operations, identifies key psychological mechanisms that mediate the effects of different collaboration modes on consumer perceptions and behavioral intentions and offers actionable guidance for service operations managers. The results suggest that GAI should be deployed strategically to augment, rather than replace, human intelligence.
Evidence weight
Balanced mode · F 0.40 / M 0.15 / V 0.05 / R 0.40
| F · citation impact | 0.50 × 0.4 = 0.20 |
| M · momentum | 0.50 × 0.15 = 0.07 |
| V · venue signal | 0.50 × 0.05 = 0.03 |
| R · text relevance † | 0.50 × 0.4 = 0.20 |
† Text relevance is estimated at 0.50 on the detail page — for your query’s actual relevance score, open this paper from a search result.