Analyzing Customer Adoption of AI Chatbots in E-Commerce: Integrating BERT-Based Sentiment Analysis with the Unified Theory of Acceptance and Use of Technology (UTAUT)
Year: 2026
Type: article
Abstract: AI-driven conversational systems have reshaped customer service operations, especially in e-commerce sites, where users expect rapid, accurate, and contextually relevant support.Despite widespread usage, the real customer acceptance of chatbots remains poorly understood due to the inherent limitations of traditional survey-based measurement approaches.This work proposes a comprehensive analytics framework which integrates BERT-based sentiment inference and the Unified Theory of Acceptance and Use of Technology (UTAUT) to evaluate the actual behavioral markers of user acceptance.In this work, a pre-trained BERT sentiment classifier will analyze a large-scale dataset of 27,000 customer support utterances.Because the dataset does not have sentiment labels, reliability is ensured via a two-fold strategy: one is literature-based benchmarking, and the other is manual human annotation.Furthermore, this work proposes, for the first time, a mathematically grounded sentiment-to-UTAUT transformation model that calculates the construct dynamics directly from the conversational signals.System architecture has been implemented to support real-time analytics with low latency.The experimental results have validated that the new method yields significant insights into Performance Expectancy, Effort Expectancy, Social Influence, and Facilitating Conditions, based not upon the survey method but upon the behavior revealed during interaction.The experimental results have shown that the sentiment-based UTAUT method can be an effective alternate way to assess the chatbot usage in the operational e-commerce setting.
Authors: Aye Chan Myae, Hlaing Htake Khaung Tin
Language: English
Topic: AI in Service Interactions
Subfield: Marketing
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W7134136299