Paper Number
ICIS2025-1180
Paper Type
Short
Abstract
Conversational breakdowns in chatbot-based customer service interactions can impair service quality and brand perception. Our study investigates how static versus generative AI (GAI)-based recovery affects user experience in such scenarios. Using a dual-study design, Study 1 examines two static strategies: Inform (explanation-focused) and Repair (action-focused). We find that Repair significantly enhances recovery quality and service perception, while Inform offers limited benefits when used alone. Study 2 explores the potential of GAI-driven recovery, employing a chatbot capable of nuanced, interactive support. Preliminary findings suggest that GAI-based recovery may increase resolution success, particularly in complex interactions, but could demand higher cognitive effort from users. By comparing both approaches and introducing a sequencing approach, our study provides novel insights into how recovery design and GAI influence service outcomes. Our results inform the design of future chatbot service systems, suggesting that adaptive, generative recovery may complement traditional static recovery strategies under specific conditions.
Recommended Citation
Benner, Dennis; Janson, Andreas; and Tan, Chee-Wee, "Less is More or More is Better? Comparing Static and Generative AI-Based Recovery for Chatbot Service Breakdowns" (2025). ICIS 2025 Proceedings. 6.
https://aisel.aisnet.org/icis2025/hti/hti/6
Less is More or More is Better? Comparing Static and Generative AI-Based Recovery for Chatbot Service Breakdowns
Conversational breakdowns in chatbot-based customer service interactions can impair service quality and brand perception. Our study investigates how static versus generative AI (GAI)-based recovery affects user experience in such scenarios. Using a dual-study design, Study 1 examines two static strategies: Inform (explanation-focused) and Repair (action-focused). We find that Repair significantly enhances recovery quality and service perception, while Inform offers limited benefits when used alone. Study 2 explores the potential of GAI-driven recovery, employing a chatbot capable of nuanced, interactive support. Preliminary findings suggest that GAI-based recovery may increase resolution success, particularly in complex interactions, but could demand higher cognitive effort from users. By comparing both approaches and introducing a sequencing approach, our study provides novel insights into how recovery design and GAI influence service outcomes. Our results inform the design of future chatbot service systems, suggesting that adaptive, generative recovery may complement traditional static recovery strategies under specific conditions.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
15-Interaction