Enhancing Veterinary Medicine with RAG-Based AI Chatbots: Improving Accessibility and Quality of Care

This thesis delves into the transformative role of AI-powered chatbots, particularly a chatbot system built on Retrieval-Augmented Generation (RAG), within the veterinary field. As digital technology progresses, the veterinary industry is seeing substantial shifts, with AI and machine learning emerging as key contributors. Smart chatbot systems create an interactive platform that offers real-time support and decision-making assistance, improving both the accessibility and overall quality of veterinary care. The proposed RAG-based chatbot utilizes AI to provide prompt, accurate, and context-aware information, helping both veterinarians and pet owners better manage animal health. The methodology section explains the design and implementation of the RAG-based chatbot. By combining both retrieval and generation approaches, it can handle large datasets and produce coherent responses. This system tackles important challenges in veterinary practice, such as ensuring quick access to trustworthy information, improving how patient data is managed, and aiding veterinarians in making well-informed decisions. Its ability to process natural language queries is enhanced by integrating Natural Language Processing (NLP) and Large Language Models (LLMs), creating an easy-to-use experience for users. Through hands-on implementation and testing, the research evaluates the system’s performance in real-world veterinary environments. The chatbot was tested on a large dataset of veterinary knowledge and received feedback from professionals in the field, demonstrating high accuracy and user satisfaction. The evaluation showed that over 856 answers were rated as 4 or higher, with notable improvements in information retrieval speed and quality of care. However, deploying AI-driven chatbots in veterinary care comes with challenges, such as ensuring the reliability of the information provided and integrating these systems smoothly into existing practices. These findings expand the current understanding of AI applications in veterinary medicine, offering strategies to overcome the hurdles associated with AI-powered solutions. Ultimately, this study aims to contribute to future developments in veterinary telemedicine, enhancing the quality of animal care and supporting veterinary professionals in their day-to-day work.