Predoc

Inspiration

Predoc was inspired by a common but critical problem in healthcare: most medical consultations are short, and patients often arrive unprepared. Symptoms are described emotionally or out of order, important details are forgotten, and clinicians must extract usable information under time pressure. This communication gap leads to frustration, repeat visits, and missed context. Predoc was built to address this gap before the consultation begins.

What We Learned

Building Predoc reinforced how important boundaries are in health-related AI. We learned that usefulness does not come from diagnosis or treatment recommendations, but from improving clarity, structure, and communication. We also learned that patients want guidance and reassurance, but health systems require neutrality, traceability, and respect for clinical authority.

How We Built It

Predoc is built with Django as the backend framework and uses a Gemini-powered AI chat system to guide patients through a structured intake conversation. The AI focuses strictly on organizing patient-reported information—such as timelines, key symptoms, relevant history, and unanswered questions—into a concise, doctor-ready brief. Safety constraints are enforced to prevent diagnosis, prescribing, or medical judgment.

Challenges

The biggest challenge was balancing usefulness with safety. We had to carefully design prompts, validation rules, and outputs to ensure the system supports clinicians rather than replacing them. Another challenge was keeping the experience simple while handling unstructured, emotional user input reliably. Solving these constraints shaped Predoc into a focused, clinically respectful tool rather than a generic health chatbot.

Built With

• django