The Impact of Visual Elements in AI Chatbots on Consumer Satisfaction

Chatbots powered by generative AI have the potential to replace many customer service jobs. While existing literature often focuses on objective performance metrics for chatbots, few research examines how chatbots could provide a more comfortable user experience. This study investigates an extension of insights from human responses to AI responses based on the SERVQUAL model, exploring how AI chatbot design can enhance consumer comfort and improve satisfaction during complaint handling. Using orthogonal arrays, we assessed seven factors in a randomized controlled trial (RCT). Results showed that human agent significantly reduced consumer satisfaction compared to non-human agents (p = 0.011), especially among male, and younger consumers. These findings indicate that when AI performance is inadequate, human-like visuals create unrealistic expectations, leading to lower satisfaction, whereas non-human visuals help better manage user expectations. Tailoring chatbot visuals elements to consumer characteristics and contexts could enhance satisfaction.