SHOPASSIST: A PERSONALIZED CUSTOMER SUPPORT CHATBOT WITH MACHINE LEARNING-POWERED INTENT RECOGNITION AND QUERY RESOLUTION

The proposed work describes the design and implementation of ShopAssist, an offline-capable customer supportchatbot system designed for e-commerce applications. ShopAssist offers a chatbot that supports customer queriesby combining semantic and contextual query understanding with user authentication and personalized customersupport. By combining TF-IDF vectorization, fuzzy string matching, and transformer-based embeddings, themodel can provide the most accurate and relevant answers in direct response to queries. With a graphical userinterface, the chatbot supports login-session-based control as well as contextual support. By combining userauthentication, contextual query understanding, and methods such as the Sentence-BERT method, Random Forestclassifiers, and Tkinter GUI elements for user interaction, ShopAssist provides offline usability in resolvingqueries. In summary, the proposed human-centred solution has immediate implications for deriving significantvalue by automating customer interactions, such as FAQs or customized responses, thereby enhancing customerengagement in a seamless digital marketplace.