FindADoctor: An AI-Powered Software for physician discovery and scheduling

UTHealth creators have utilized artificial intelligence (AI) physician discovery and scheduling platform built on the Model Context Protocol (MCP), where patients can always see who is available right now, using a hybrid architecture that ensures the most clinically appropriate and immediately accessible providers appear first, cross-referencing department-level Areas of Interest, subspecialty scopes, patient eligibility, and live scheduling data to produce a single, prioritized list easily navigated by patients.

Current Challenges with Finding a Physician

Finding the right doctor today can be unnecessarily difficult: patients scroll through outdated directories, call overburdened scheduling lines, and guess at which provider actually treats their condition, speaks their language, and has an opening this week. Traditional "find a doctor" tools rely on static, batch-loaded databases that grow stale between refresh cycles, leading to inaccurate availability, missing specialties, and poor search relevance.

Introducing: FindADoctor

UTHealth Houston creators have developed FindADoctor, an AI-powered physician discovery and scheduling platform built on a hybrid architecture, combining MCP with GPT-5 to interpret natural language, while cross-referencing department-level interests, subspecialty scopes, patient eligibility (age, insurance), and live scheduling data to produce a single, prioritized list that the patient can easily navigate.

Technology Differentiators and Potential Benefits

Features that make FindADoctor unique include:

• Connecting patients to the right provider through GPT-5 natural language search, i.e. interpreting queries like "thyroid surgeon in Sugar Land" or "pediatrician who speaks Spanish" and matching them to physicians in real time using live data from Epic.

• Leverages a remote MCP server that acts as a real-time bridge to the EHR provider directory, returning authoritative, domain-ranked results with live appointment availability—eliminating the reliance on stale, batch-processed data common in traditional physician finders.

• Inclusion of a hybrid AI engine powered by LLM extracts structured intent (specialty, location, language, age group) from free-text queries that applies intelligent re-ranking using clinical metadata, condition indices, and geographic matching to surface the most relevant providers first.

• Delivering on-demand appointment availability directly from an EHR web widget through the MCP layer, enabling patients to see next-available slots and schedule online in a single, seamless interaction.

Intellectual Property Status

Available for licensing

About the Creators

Martin J. Citardi, MD

Vice Dean for Clinical Technology

Chair of Otorhinolaryngology-Head & Neck Surgery

Xiaoqian Jiang, Ph.D.

Associate VP for Medical AI, Chair of Department of Health Data Science and Artificial Intelligence

Professor in Biomedical Informatics and Bioengineering

Jiajie Zhang, Ph.D.

Dean, Professor, Chair of McWilliams School of Biomedical Informatics