ββββββββββββ βββββββββββ βββββββ βββ ββββββ ββββββββββββββββββ
βββββββββββββ βββββββββββββββββββββββββ ββββββ βββββββββββββββββββ
ββββββ ββββββββββββββββββββββ ββββββ ββ βββββββββββββββββ ββββββββ
ββββββ ββββββββββββββββββ βββ βββββββββββββββββββββββββββ ββββββββ
βββββββββββ βββ ββββββ ββββββββββββββββββββββ ββββββββββββββ βββ
βββββββββββ ββββββ βββββββ ββββββββ βββ ββββββββββββββ βββ
π Live Demo β’ π₯ Problem β’ β¨ Features β’ ποΈ Tech β’ π Impact
What if every doctorβs recommendation automatically turned into completed care?
In gynecology, the hardest part of treatment often isnβt the diagnosis-itβs the follow-through. Patients leave appointments with unclear instructions, referrals get delayed, and administrative bottlenecks slow down care. empowHER solves this problem. Our AI-powered workflow engine listens to clinical conversations, extracts key action items, and automatically generates patient-ready care plans, symptom tracking tools, and complete referral packets. By converting unstructured visits into structured workflows, empowHER ensures that clinical decisions donβt stop at the chart-they move patients forward.
|
Time on EHR vs Time with Patients Physicians spend nearly Annals of Internal Medicine, 2016 |
Referrals That Never Complete Over 50 million Patients fall through cracks AAFP, 2018 |
Patients Who Don't Understand Only 13% truly comprehend Medical jargon kills Academic Emergency Medicine, 2012 |
graph TD
A[π¨ββοΈ Physician drowning in paperwork] -->|49% of day| B[β οΈ Less time with patients]
B --> C[π Patient leaves confused]
C --> D[π Care plan misunderstood]
D --> E[β 40% non-adherence]
E --> F[π Preventable complications]
F --> G[π° Expensive emergencies]
A --> H[π Referral needed]
H --> I[π Paperwork friction]
I --> J[β 50% never complete]
J --> G
style A fill:#ff6b6b
style G fill:#ff0000
style J fill:#ff0000
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β β
β ποΈ AUDIO β π€ AI β π STRUCTURE β β¨ ACTION β
β β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
|
LISTEN |
UNDERSTAND |
STRUCTURE |
DELIVER |
|
The Reality: Doctors spend twice as long typing as they do listening to patients The Fix: AI that captures everything, structures everything, routes everything
45 minutes per day reclaimed per clinician |
The Reality: 78% of patients walk out with no idea what to do next The Fix: AI that translates "medical" into "human"
Comprehension: 13% β 85% |
|
The Reality: Patients forget. Doctors ask the same questions twice. 40% don't follow through. The Fix: A longitudinal memory that builds the story
3x higher patient engagement |
The Reality: Half of all referrals die in transit. Paperwork. Confusion. Missing forms. The Fix: Coordination that guarantees completion
Completion: 50% β 90% |
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β π€ INPUT LAYER β
β Audio β’ Documents β’ Symptoms β’ Benefits β’ Referral Requests β
βββββββββββββββββββββββββββββ¬ββββββββββββββββββββββββββββββββββββββ
β
βββββββββββββββββββββββββββββΌββββββββββββββββββββββββββββββββββββββ
β β‘ FRONTEND β Next.js 14 β
β TypeScript β’ Tailwind CSS β’ shadcn/ui β’ Recharts β’ React PDF β
β β
β ββββββββββββββββ ββββββββββββββββ ββββββββββββββββ β
β β Clinic β β Patient β β Charts β β
β β Dashboard β β Portal β β & Analytics β β
β ββββββββββββββββ ββββββββββββββββ ββββββββββββββββ β
βββββββββββββββββββββββββββββ¬ββββββββββββββββββββββββββββββββββββββ
β REST API
βββββββββββββββββββββββββββββΌββββββββββββββββββββββββββββββββββββββ
β π§ BACKEND β FastAPI (Python) β
β Redis β’ Celery β’ Authentication β’ Orchestration β
βββββββββββββββββββββββββββββ¬ββββββββββββββββββββββββββββββββββββββ
β
βββββββββββββββββββΌββββββββββββββββββ¬ββββββββββββββ
β β β β
βββββββββββΌβββββββββ ββββββββΌβββββββ ββββββββββΌββββββββ βββΌβββββββββββ
β π€ Watson STT β β π§ watsonx β β πΎ IBM Db2 β β π¦ Object β
β β β .ai β β β β Storage β
β β’ Medical vocab β β β’ Reasoning β β β’ Visit data β β β’ Audio β
β β’ Real-time β β β’ Summaries β β β’ Symptoms β β β’ PDFs β
β β’ Multi-speaker β β β’ Extractionβ β β’ Referrals β β β’ Packets β
ββββββββββββββββββββ βββββββββββββββ ββββββββββββββββββ ββββββββββββββ
β
βββββββββββββββββββββββββββββΌββββββββββββββββββββββββββββββββββββββ
β π€ OUTPUT LAYER β
β Care Plans β’ Summaries β’ Packets β’ Notifications β’ Analytics β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
|
|
|
|
|
| β Traditional EMR | β EmpowHER |
|---|---|
| Doctors type while patients talk | AI listens, doctors connect |
| Generic summaries nobody reads | Personalized plans patients follow |
| Symptoms forgotten between visits | Continuous tracking builds the story |
| Referrals lost to paperwork | Automated handoffs with zero friction |
| Providers drowning in admin | 60% less burden, better outcomes |
| Optimized for billing | Optimized for healing |
|
Every claim backed by Annals of Internal Medicine Not guesses. Facts. |
Not a hackathon toy Enterprise architecture Ready for 1,000 clinics |
Sun Life: Women's health Direct alignment No guessing needed |
# Clone the magic
git clone https://github.com/your-team/empowher.git
cd empowher
# Frontend setup
npm install
cp .env.example .env.local
# Add your IBM credentials to .env.local
# Backend setup
cd backend
pip install -r requirements.txt
cp .env.example .env
# Add your IBM credentials to .env
# Fire it up
npm run dev # Frontend β http://localhost:3000
uvicorn app.main:app --reload # Backend β http://localhost:8000
# Watch the magic happen π©β¨# IBM Cloud (Required)
IBM_WATSON_API_KEY=your_watson_key
IBM_WATSON_URL=your_watson_url
WATSONX_PROJECT_ID=your_project_id
DB2_CONNECTION_STRING=your_db2_string
IBM_COS_API_KEY=your_cos_key
# App Config
NEXT_PUBLIC_API_URL=http://localhost:8000|
Frontend/UX Lead Next.js Building the interface |
Workflow/API Lead FastAPI Making the backend |
AI Integration Lead watsonx.ai Teaching AI |
Data/Security Lead Db2 Keeping patient data |
| Security Layer | Implementation |
|---|---|
| π Encryption at rest | AES-256 (Db2 + Object Storage) |
| π Encryption in transit | TLS 1.3 for all API calls |
| π€ Access control | IBM Cloud IAM with role-based permissions |
| π Audit logging | Every PHI access tracked in Db2 |
| β Patient consent | Explicit approval for audio recording |
| π« No diagnosis | Supportive tool, not medical device |
|
|
Healthcare where information flows seamlessly,
patients understand completely,
and clinicians focus on healing instead of paperwork.
That's not a fantasy. That's the system we're building.
MIT License β Built for GenAI Genesis 2026