Optimizing the Design of a Business English Multilingual Customer Service System Using Clustering Algorithms

Globalized industries increasingly require multilingual customer service systems capable of accurately processing specialized technical terminology. This demand is particularly significant in electromagnetic and antenna-related industries, where communication involving antenna specifications, electromagnetic compatibility, wireless system parameters, and technical service requirements requires high semantic precision across languages. To improve the understanding and classification of multilingual Business English queries, this study proposes an optimized customer service framework based on hybrid clustering and multilingual BERT (mBERT) embeddings. The framework consists of three components: language detection and preprocessing, semantic embedding generation using mBERT combined with TF-IDF weighting, and a hybrid clustering mechanism integrating K-means with semantic similarity refinement. Experimental results demonstrate that the proposed approach improves clustering quality, classification accuracy, and processing efficiency for multilingual technical service scenarios. The framework provides an effective solution for intelligent customer support involving specialized technical communication in international business environments.