Publication

Abstract : Prompt Engineering deals with the design, refinement and implementation of prompts. The new research field guides the response of large language models to perform various tasks. Recent studies show significant emergence of chatbots for numerous e-commerce applications. Prevailing chat platforms are characterized with static response. Recently, chatbot technology has experienced a tremendous advancement with the incorporation of ChatGPT. This paper introduces AidBot, a new customer care ChatGPT based chatbot that responds dynamically by providing human like response to the customers. The AidBot, trained deploying Chain of thought process, encompass frontend for a lively conversation and a backend that embeds OpenAI API and a storage to log the conversation details. Unlike present chatbots, AidBot uses prompt engineering technology to make the conversation more human-like. The proposed system is tested for various conversations that includes type grammatical mistakes to analyse its response. Moreover, the adaptability of the model to multiple languages are also tested. The effectiveness of the model in storing the required conversation details and sharing with the associated customer care officials are tested. The suggested system is expected to provide a user friendly and convincing customer service based on the past conversations.