AI-Driven Complaint Management for Rail

Abstract

Millions of passengers use India's railways every day, which leads to a high volume of complaints about food, cleanliness, coach conditions, safety, and timeliness. Passengers become dissatisfied with traditional complaint redressal systems because they are frequently manual, slow, and opaque. Our solution to these problems is an AI-Driven Complaint Management System that uses OCR (Optical Character Recognition), Machine Learning, and Natural Language Processing (NLP) to automate the classification, prioritization, and tracking of complaints' resolutions. OCR processes complaints that are submitted as text, photos, or videos and classifies them into the relevant departments (e.g., cleanliness, AC, food, safety). TF-IDF + Logistic Regression models are used to determine priority levels (high, medium, and low), and optional LLM integration improves semantic understanding. Streamlit was used to create an admin dashboard that offers real-time analytics, complaint tracking, and resolution management. This system can be implemented in smart railway operations since it increases passenger satisfaction, efficiency, and transparency.