Utilizing AI to manage rapport: ERNIE Bot’s interpersonal strategies to respond to negative hotel reviews

Abstract

Large language models have exhibited satisfactory performance in linguistic tasks pertaining to lexicon, grammar, and syntax. However, their pragmatic competence in user interactions remains underexplored. To address this research gap, the current study delves into the interpersonal dimension of pragmatic competence, specifically focusing on ERNIE Bot’s potential to manage rapport with reviewers in response to negative hotel reviews on TripAdvisor. A comparative approach is employed to analyze the rapport-managing moves in responses generated by ERNIE Bot versus those authored by human hotel managers. Our findings indicate that ERNIE Bot frequently employs rapport-enhancing moves to foster relationships. Specifically, we identify three key interpersonal strategies used by ERNIE Bot: specifying concerns, supporting emotions, and undertaking responsibility. The results suggest that ERNIE Bot demonstrates its pragmatic competence to promote positive rapport with reviewers. However, further training is necessary to align its responses more closely with hotels’ transactional objectives. Overall, this study provides insights into large language models’ linguistic potential in interpersonal contexts and offers practical recommendations for professionals considering the integration of AI-generated content into their operations.