Building a Task-Oriented AI Assistant for a Smart Bus Booking System

Tóm tắt

Abstract - In the context of robust digital transformation, Vietnam's passenger transport industry faces an urgent need to modernize processes and enhance customer experience. Traditional booking methods, such as telephone calls or in-person counter purchases, are revealing numerous limitations: they are time-consuming, lack price transparency, and involve manual ticketing processes prone to errors. This paper proposes the development of a multi-platform (Web and Mobile) smart bus booking system to address these challenges. The system is built upon a modern microservice architecture. Specifically, the backend services are developed using Node.js, while the frontend layer utilizes ReactJS for the Web platform and React Native for Mobile. This architecture also integrates Redis as a solution for data caching and session management, helping to reduce the query load on the primary database. The system's core innovation is the integration of an AI Assistant powered by the Gemini model, applying a Retrieval-Augmented Generation (RAG) architecture. Instead of providing generic responses, this AI Assistant is capable of understanding natural language and directly querying internal microservices (such as the Trip & Schedule Service and Promotion Service). This allows it to provide accurate information and automate the booking process, thereby minimizing time and enhancing user satisfaction.