BeforeYouGo

BeforeYouGo is an AI-powered healthcare assistant that turns your daily health data into insights and preps you for doctor visits

Comment

Inspiration

Healthcare is fragmented, reactive, and hard to navigate, especially for patients managing multiple data points like symptoms, appointments, insurance, and medications.

We noticed that people often walk into doctor appointments unprepared, without clear context about their own health.

We wanted to build a system that helps users understand their health, organize critical information, and make better decisions before they even see a doctor.

What it does

BeforeYouGo is an AI-powered healthcare assistant that helps users prepare for and manage their medical care.

It:

  • Tracks health data (wearable + manual input)
  • Generates AI-driven health summaries and insights
  • Helps users log symptoms, mood, and daily activity
  • Prepares users for doctor visits with personalized questions and talking points
  • Organizes appointments, medications, and insurance information in one place
  • Provides an AI chat assistant for health-related guidance

The goal is simple:

turn scattered health data into actionable, personalized guidance.

How we built it

  • Frontend: Xcode (iOS app)
  • Backend: Laravel (API + data management)
  • Database: Structured user health + profile data storage
  • AI Layer: RAG used by LLMs for:
    • Health summaries
    • Appointment preparation
    • Post-visit analysis
  • Integrations:
    • HealthKit for real-time health data
    • Camera input for profile picture capture
  • Architecture:
    • Modular system separating health tracking, AI processing, and user workflows
    • Real-time sync between frontend and backend

Challenges we ran into

The biggest technical challenge was OCR for insurance card scanning.

We initially attempted to extract structured data (provider, member ID, etc.) directly from images, but ran into:

  • Inconsistent formatting across insurance cards
  • Low accuracy in extraction without heavy preprocessing
  • Time constraints that made refining OCR unreliable

Instead of forcing a weak solution, we made the decision to remove OCR and pivot to manual input, ensuring reliability and a better user experience.

We also faced challenges in:

  • Structuring health data in a way that AI could meaningfully interpret
  • Balancing feature scope with time constraints
  • Keeping the UI clean despite a large feature set

Accomplishments that we're proud of

  • Built a fully integrated healthcare assistant in a short time frame
  • Successfully combined health tracking + AI insights + appointment finder & prep into one system
  • Designed a workflow that actually mirrors real-world healthcare behavior
  • Delivered a functional, end-to-end product instead of isolated features
  • Made a strong product decision by cutting OCR instead of shipping something unreliable

What we learned

  • In healthcare, accuracy is more important than feature count
  • AI is powerful, BUT only when paired with structured, meaningful data
  • Trying to do too much can kill execution, prioritization matters
  • Real-world systems require handling messy, inconsistent data!
  • It’s better to ship a smaller, reliable system than a larger, fragile one

What's next for BeforeYouGo

  • Revisit OCR with a more robust pipeline (better models + preprocessing)
  • Add deeper personalization using long-term health trends
  • Integrate with more health data sources and providers
  • Expand AI capabilities for early risk detection and recommendations
  • Improve provider matching and insurance compatibility features
  • Move toward a predictive healthcare assistant, not just a reactive tool

Built With

Share this project:

Updates