Trust Formation in AI-Powered Customer Service Systems: A Longitudinal Study

Abstract

AI-powered chat systems are widely adopted in customer service platforms to reduce response time and operational cost. However, sustained user acceptance depends on more than initial convenience. This paper investigates how users develop trust toward automated support agents during repeated service interactions. Special attention is given to error recovery behavior, explanation quality, and escalation mechanisms. Observations from long-term platform deployment suggest that transparency and predictable fallback procedures play a greater role in trust formation than conversational fluency alone.