AI-Powered Chatbots in Public Services: The Case of PostQuest for Government Schemes
摘要
Artificial intelligence (AI) and machine learning (ML) are tremendously changing the lives of people across the globe. The chatbot can be a powerful medium to bridge the gap between common people and technology to access information about different government schemes. With this motivation, the current communication introduces an innovative AI chatbot in Indian Languages to enhance user interaction and information accessibility. One of the widely used government public services in India is Post Bank Services. This study presents an AI-powered chatbot ‘PostQuest’ to assist common people in accessing Indian Post Bank Schemes details. Utilizing cutting-edge natural language processing (NLP) models such as Mistral-7B-Instruct-v0.1 and Llama-2-13B-Chat, the PostQuest chatbot is trained on a dataset of 1000 questions to ensure personalized assistance with accurate and contextually relevant response in English and five major Indian languages. Key technological components include voice recognition, text conversion modules, and a robust NLP translator. These technologies work together to process user queries and deliver responses in text and speech formats, supporting multilingual communication. The system integrates web development frameworks for a seamless user experience that is accessible via web applications. With an accuracy rate of 67.45% in generating semantically similar responses, the chatbot demonstrates the effectiveness of AIML learning in improving information dissemination and user engagement. Using Fuzzy Wuzzy with a threshold of 85, the string matching has accuracy of 97.67%. Its multilingual support broadens accessibility, making it a valuable tool for diverse user bases in India and enhancing overall user interaction and satisfaction.