Our medical AI chatbot strives to make a user-friendly experience for students who want to request an appointment, get a pricing estimate, or obtain medical information with little-to-no hassle.
We noticed that the website for UVA's health department was somewhat confusing and many of the pages were redundant.
What this chatbot was made for was to eliminate these redundancies and make a smoother process when trying to do common tasks such as scheduling an appointment, getting a pricing estimate for medical services or products, or simply obtaining general information regarding the UVA health department.
We developed this chatbot by importing the OpenAI API into our Django project for the backend in order to use LLMs and SQL, while using React and the webdev stack for the frontend. Additionally, we utilized Vite for live optimized web building and dynamic user-friendly interaction.
We ran into the challenge of figuring out how to exactly implement the AI aspect of the chatbot, as well as researching the appropriate usages of LLMs.
Despite this being our first experience developing with AI, we managed to make a solid chatbot that gives us ample and respective information from the content and information that we've given it.
We learned how to style frontend components using React, HTML, and CSS, as well as how to train an AI chatbot using databases and information that we supply to it.
Although we completed our general goal of making a helpful AI chatbot, we could not reach the full potential as the databases we hoped to use were not accessible. For example, we could not give truly accurate information regarding pricing estimates since we cannot access insurance rates specific to users and companies. Additionally, since we do not have the appointment scheduling database for UVA health centers, we could not actually create an appointment.
Log in or sign up for Devpost to join the conversation.