This project introduces an AI-powered automation system designed to streamline booking workflows for small and medium-sized enterprises (SMEs) in the car rental industry. Utilizing a Telegram chatbot interface, the platform leverages a fine-tuned BERT (Bidirectional Encoder Representations from Transformers) model for high-accuracy intent classification. The system automates key operational tasks, including query handling, real-time availability checks, and structured data logging to a Google Sheets backend. Experimental results demonstrate that the BERT-based Natural Language Understanding (NLU) module achieves a validation accuracy of approximately 94%, effectively classifying core booking intents even with varied natural language inputs. By replacing traditional manual processes with this lightweight, scalable architecture, the system reduces operational effort by an estimated 80-90%, offering a cost-effective and accessible automation solution for businesses with limited technical infrastructure.
Kaushal Shete (Sun,) studied this question.