AN AI-POWERED E-COMMERCE PLATFORM WITH SEMANTIC SEARCH, AI PRODUCT INFORMATION GENERATION, SMART PRODUCT COMPARISON, AND CONVERSATIONAL CHATBOT SUPPORT.

Abstract

This thesis presents the design and implementation of an AI-powered e-commerce platform for electronics retail. The system integrates traditional online shopping features with artificial intelligence capabilities, including semantic product search, automated product information generation, review-based question answering, personalized recommendations, and a conversational chatbot. The platform is built using Spring Boot 3 for the backend, Vue 3 for the frontend, and PostgreSQL with the pgvector extension for vector similarity search. The methodology combines embeddings, hybrid recommendation algorithms, and retrieval-augmented generation to bridge the gap between user intent and structured product data. A set of evaluation scripts is used to measure search relevance, recommendation quality, and chatbot response usefulness on curated test cases. Results show that AI features can enhance product discoverability and decision support, while the discussion highlights current limitations in data volume, language support, and scalability, and proposes directions for future improvement.