Designing an AI-powered health ecosystem from scratch — where doctors consult smarter and patients carry their complete medical story in one tap.
Healthcare suffers from a painful paradox: the tools meant to help are slowing everyone down. Every consultation involves repetitive prescription writing, manual record-keeping, and hours spent deciphering scattered patient histories.
On the patient side, medical records live across hospitals, clinics, and lab folders — making it nearly impossible to have a unified health view when it matters most.
For Doctors: Repetitive prescription writing, manually analyzing patient history, no efficient way to track health trends over time.
For Patients: Records lost across clinics. No way to share complete history with new doctors. Zero visibility into their own health patterns.
For Both: Communication gaps between consultations lead to missed follow-ups, forgotten medication changes, and fragmented care.
What if a doctor could understand a patient's entire health story in seconds — and patients could carry their complete medical history in their pocket?
Created the complete brand from zero — the sefit logo with its inline typeface and heart+medical cross symbol, the Denim & Selective Yellow color palette, Source Serif Pro + Inter typography, and a soft 3D imagery direction.
Mapped user journeys for doctors (web) and patients (mobile), designed the timeline view, AI recorder flows, access code system, and 56+ screens across both platforms in Figma.
Built interactive Figma prototypes, tested concepts using vibe coding tools like Lovable for rapid validation, and coordinated with manager, developers, and client for seamless handoffs.
Deep stakeholder interviews to understand doctor workflows and patient pain points. Defined the information architecture — how records flow between patient app and doctor dashboard, how AI categorization works, and the access code permission system.
Designed the brand from scratch: the custom inline logo with heart + medical cross, the Denim blue and Selective Yellow palette, Source Serif Pro for titles paired with Inter for body text, and a soft 3D imagery style blending warmth with medical precision.
Built the web dashboard — patient listing, timeline records with filters, AI recorder, AI summary generation, trends, messaging, and Ask AI. Simultaneously designed the mobile app — onboarding, camera scan, timeline view, family profiles, doctor listing, chat, and access code sharing.
Built interactive Figma prototypes for both platforms, refined edge cases and empty states, iterated on developer and client feedback, and delivered production-ready design documentation.
Denim blue for trust, Oxford Blue for depth, Selective Yellow for warmth and focus — a palette that blends medical authority with modern AI innovation.
A dashboard that surfaces patient insights in seconds. Instead of digging through files, doctors get a single intelligent interface — with AI handling prescriptions, summaries, and records — so they can focus on the patient in front of them.
The flagship feature. Doctors hit record — the AI captures the conversation, extracts medical data, and auto-generates a structured record with diagnosis, prescriptions, and follow-up advice. Three input paths: upload document, AI recorder, or manual entry — giving doctors full flexibility.
Patients organized by today's appointments with search. Records locked by default — doctors enter a patient's unique access code to unlock. Privacy-first by design.
Vitals, ongoing prescriptions, infections, cardiac evaluation, chronic conditions — all AI-generated from records. Trends tab shows visual charts over time.
Direct chat with patients for follow-ups, medication queries, and ongoing care. Messages sit alongside records in the patient profile.
AI assistant that summarizes clinical findings, analyzes records, and offers suggestions. Plus a built-in appointment scheduler with clinician and patient assignment.
Interactive Prototype — 21 Web Screens
Explore Web Prototype →Patients snap a photo of any medical document — and the AI categorizes it, builds a timeline, generates summaries, and makes it shareable with any doctor via a unique access code. One app becomes the family's health vault.
Upload via camera scan, file upload, or AI voice recorder — the AI categorizes documents (Prescription, ECG, Discharge Summary, Blood Test, X-Ray, Consultation Note) and arranges by date into a visual timeline. Filter chips let you view by type. What once took hours happens in one tap.
AI creates a living health summary from records — vitals, ongoing prescriptions, infections, cardiac evaluation, chronic conditions. Doctors review this instead of reading every report.
Switch between profiles (Myself, Mother, Wife, Son) from the home screen. Each gets their own timeline, summary, and records. One app for the whole family.
Browse doctors with specialty filters (MBBS, Cardiologist, Orthopedic), connect via in-app chat, and book appointments by calling the hospital (Phase 1).
Share records via a unique 4-digit code — controlled, toggleable per doctor, and lockable with one tap. See who has access and revoke anytime. Privacy by design.
Interactive Prototype — 35 Mobile Screens
Explore Mobile Prototype →Doctors have seconds, not minutes. Every screen had to pass the "can a doctor do this between patients?" test. This pushed me toward information-dense but scannable layouts — like the timeline with filter chips and the summary tab's structured sections.
The AI features are powerful. But without clear labels like "This summary is AI-generated — always review for accuracy," doctors won't adopt them. I designed AI as a collaborator with approval flows and edit capabilities — never a black box.
Designing for doctors (web) and patients (mobile) simultaneously meant thinking systemically. The access code system, shared record format, and mirrored timeline — both platforms feel like one product despite different mental models.
56+ screens, complete brand identity, interactive prototypes — all in 4 weeks. Using Lovable for rapid prototyping and maintaining a tight feedback loop with developers and client made it possible without sacrificing quality.