Conversational AI-based Smart Ticketing System for Public Transit using Machine Learning

There are general challenges in the usability of rail ticket booking especially with first time, old and visually challenged users. This paper presents a conversational AI-driven train ticket system that can make booking easier using a chatbot combining both text and voice interaction. The system utilizes TF-IDF as well as K-Nearest Neighbors (KNN) to classify intents whilst a Random Forest Regressor predicts fares according to dynamical travel parameters. The application is implemented with React.js, FastAPI, and SQLite and supports the protection of e-ticket generating and OTP-based authentication processes. The analysis conducted experimentally proves that intent recognition is accurate, the fare calculation is effective, and the accessibility is enhanced with different users. The results indicate the possibility of improving the inclusiveness, efficiency, and usability of the current transport ticketing system by using conversational AI.