Recovering trust through Showing Competence & Building Rapport: You Need to Provide Corrective Answer in Formal Tone
Preprint
- 29 April 2026
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
GenAI is increasingly used for information-seeking tasks, yet GenAI service failure remain unavoidable, making trust recovery essential for sustaining user engagement. Drawing on justice theory and communication accommodation theory, this study examines how recovery strategies and tone of voice jointly regain trust resilience, which in turn fosters loyalty. Using a scenario experiment, we found that providing corrected answers (utilitarian recovery) more effectively rebuilds trust (trust resilience) than apologies alone (symbolic recovery), and that a formal tone outperforms an informal tone in the relationship repair stage. Perceived anthropomorphism further moderates these effects by strengthening competence and altering tone-related relational responses. The findings introduce trust resilience as a key post-failure mechanism, clarify the central role of substantive corrective action, extend communication accommodation theory to trust repair, and offer practical guidance for designing GenAI recovery strategies that balance informational accuracy with relational sensitivity.Keywords
This publication has 97 references indexed in Scilit:
- Academic writing and ChatGPT: Students transitioning into college in the shadow of the COVID-19 pandemicDiscover Education, 2024
- Effects of communication style on relational outcomes in interactions between customers and embodied conversational agentsPsychology & Marketing, 2023
- The dark sides of AI personal assistant: effects of service failure on user continuance intentionElectronic Markets, 2021
- Understanding anthropomorphism in service provision: a meta-analysis of physical robots, chatbots, and other AIJournal of the Academy of Marketing Science, 2021
- Eliza in the uncanny valley: anthropomorphizing consumer robots increases their perceived warmth but decreases likingMarketing Letters, 2019
- Detecting experts on Quora: by their activity, quality of answers, linguistic characteristics and temporal behaviorsSocial Network Analysis and Mining, 2015
- A new criterion for assessing discriminant validity in variance-based structural equation modelingJournal of the Academy of Marketing Science, 2014
- The nature of an apology: An experimental study on how to apologize after a service failureMarketing Letters, 2012
- e‐WOM Scale: word‐of‐mouth measurement scale for e‐services contextCanadian Journal of Administrative Sciences / Revue Canadienne des Sciences de l'Administration, 2010
- Machines, social attributions, and ethopoeia: performance assessments of computers subsequent to "self-" or "other-" evaluationsInternational Journal of Human-Computer Studies, 1994