AI Medical Appointment Assistant: An Intelligent Hospital Registration System Based on LLM and RAG

An intelligent conversational AI medical appointment assistant designed for Amrita Hospital, combining large language models (LLM) and retrieval-augmented generation (RAG) technology to automate medical appointment processing and provide intelligent responses to patient inquiries. The goal is to enhance healthcare service efficiency and patient experience. The project is developed and maintained by Arjunch2003, open-sourced on GitHub, and released on June 15, 2026.

Traditional medical appointments have many pain points: waiting on hold during peak phone appointment times, high barriers to using online systems, and manual consultations consuming a lot of medical staff resources. To address these issues, this project builds an intelligent conversational system that combines LLM's semantic understanding capabilities with RAG's knowledge base querying to provide a smooth conversational experience and accurate, professional responses.

LLM-Driven

Uses large language models to understand patients' natural language input, process non-standard expressions, and accurately identify real intentions (e.g., recognizing the need for neurology from "headache").

RAG Technology

Combines external knowledge base retrieval: builds structured knowledge bases for hospital doctor information, department schedules, etc. → understands queries → retrieves relevant documents → generates answers based on real data to avoid model hallucinations.

Conversation Management

Maintains multi-turn conversation context, tracks appointment process steps, and ensures complete information collection.

1. Intelligent Department Recommendation: Recommends departments based on symptoms (e.g., recommending cardiology/respiratory medicine for "chest tightness");
2. Doctor Matching and Appointment: Displays a list of department doctors (specialties, available appointment times, reviews) and supports appointment completion via conversational selection;
3. Intelligent Q&A: Answers official hospital information such as consultation processes, preparation items, fees, and transportation;
4. Multi-turn Conversation Support: E.g., guiding patients through the complete appointment process: department selection → doctor selection → time confirmation.

For Patients

• Available 24/7 for appointment and consultation anytime, anywhere;
• Zero waiting, instant response;
• Natural language interaction with low operation barriers;
• Rich information query to prepare for consultations.

For Hospitals

• Reduces manual pressure and frees up medical staff time;
• Improves appointment efficiency and reduces patient decision-making difficulties;
• Accumulates data to optimize services;
• Standardizes services to ensure consistent and accurate information.

The current prototype needs improvement in the following areas:

• Strengthen medical safety boundaries to prevent providing diagnostic advice;
• Add multilingual support;
• Integrate with the hospital's HIS system to enable real-time appointment slot query and confirmation;
• Ensure compliant storage of patient data privacy. This project provides a valuable reference implementation for smart healthcare.