← Back to portfolio
Telemedicine · AI Health · Pharmacy · Labs

A nation's
healthcare,
in one tap.

One system from a 2am symptom to medicine at the door — AI triage, video consults, e-prescriptions, pharmacy, labs and emergency response, across patient, doctor and admin.

Role
Lead Product Designer
Scope
Research → UX → UI → Test
Timeline
9 months
Platforms
iOS · Android · Web
HealthTechTelemedicineAI TriageDesign Systems0→1 ProductMobile + Web
OnboardingDashboardHealth warning
OnboardingDashboardHealth warning
AI Triage
Low risk · self-care
Upcoming
Dr. Sarah · 10:30 AM
Vitals
BP 120/80
Pharmacy
Medicine delivered ✓
Rated
4.8 · 12.4k reviews
01 — Project Overview

One ecosystem. Four products. Three roles.
Designed to behave as a single organism.

A patient app, doctor app, clinician portal and admin console — built to act as one system, not four.

Patient App

0 screens

Triage, consults, pharmacy, labs, records & SOS.

Doctor App

0 screens

Schedule, video visits, prescribing & earnings.

Doctor Portal

0 screens

Onboarding, verification, documentation, payouts.

Admin Console

0 screens

Network ops, compliance, pharmacy, labs, analytics.

0+
Screens designed end-to-end
0
Core flows mapped
0
AI capabilities built in
0
Design-system components
02 — Executive Summary

Fragmented care, made continuous.

One record carries the patient from a 2am symptom to a resolved outcome — six moves that never leave the platform.

The journey · problem → impact
01
Problem
Slow, fragmented, low-trust care
02
Research
32 interviews · 410 surveys
03
Insight
Trust & continuity win
04
Strategy
AI front door + care graph
05
Solution
4 products, 1 record
06
Impact
+38% conv · 91% complete · 4.8★
We didn't design a healthcare app — we designed the connective tissue between every person, doctor and pill in the system.
03 — Problem Space

The care journey leaks
at every handoff.

A person with a symptom passes through five disconnected systems — and most give up before treatment ever starts. Each drop-off is a person who stayed sick, and revenue that never landed.

0h
Median wait for a routine appointment
1 in 0
Abandon when meds/labs live in another app
0%
Distrust online doctors without proof
0%
Of a doctor's day lost to admin
Where today's journey breaks

100 people notice a symptom. By the time treatment begins, two-thirds are gone.

100
Symptom
64
Search care
47
Book
38
Consult
33
Treated

Patients

  • Don't know what's urgent
  • Records scattered everywhere
  • Repeat their history each time

Doctors

  • Admin eats clinical time
  • No context before a consult
  • Opaque, slow payouts

Operations

  • Manual doctor verification
  • Reactive to incidents
  • No single source of truth

Business

  • Leaky booking funnel
  • Single-use, low retention
  • No moat beyond price
A person with a 2am fever shouldn't have to choose between guessing on Google and a $400 ER bill.
04 — Market Opportunity

A category compounding at double digits.

The winner won't be a point tool — it'll be the platform that owns the connected journey, with trust and AI as the wedge.

▲ Market
$0B
Digital health by 2030
▲ CAGR
0%
Telemedicine growth
▲ Adoption
0%
Used telehealth this year
▲ AI
0%
Health systems piloting AI
Telehealth adoption, % of adults
11
'19
32
'20
29
'22
35
'24
38
'25
47
'28e

The gap isn't access to doctors. It's access to continuity — a system that remembers you.

Aging populationsClinician shortagesRural access gapsCost pressureConsumerized health
05 — Research & Discovery

We listened before we wireframed.

Mixed methods built to separate what people say from what they actually do.

0
In-depth interviews
patients · doctors · admins
0
Survey respondents
5 regions
0
Usability sessions
2 rounds, moderated
0
Competitor screens audited
4 products

What the data clustered into

0%

delay care — booking feels like a chore

0%

want "is this serious?" before booking

0%

distrust doctors without verification cues

3.2×

more likely to return if records persist

⅔

abandon when fulfillment is separate

06 — Competitive Analysis

Strong players. Obvious white space.

Each rival owns a slice — booking, telehealth or AI checks. None owns the connected loop from triage to delivery with trust built in.

CapabilityPractoApollo 24/7TeladocAdaMediCare+
AI symptom triage
Video consultation
Integrated pharmacy
Lab booking + AI reads
Unified health record
Emergency / SOS

Strong   Partial   Absent

Strengths

Trust, doctor supply, brand recall.

Weaknesses

Siloed features, clunky handoffs.

Gap

No AI front door tied to fulfillment.

Our opening

Own the connected loop, trust-first.

07 — User Personas

Three people. One system serving them all.

A patient, a doctor and an admin — different stakes, one shared record.

A

Alex, 34

Busy patient · urban
Goal

Fast answers, trusted doctors, meds without a pharmacy trip.

Friction

Can't tell what's urgent; hates re-explaining history.

"Tell me if it's serious, then handle the rest."
S

Dr. Sarah, 41

Cardiologist · panel of 600
Goal

See more patients with less admin; predictable income.

Friction

Tool-switching; no pre-visit context.

"Brief me, let me prescribe, pay me on time."
M

Maya, 38

Network admin · operations
Goal

A safe, compliant, high-utilization network.

Friction

Manual verification; reactive to incidents.

"One console to see, approve, and act."
08 — Jobs To Be Done

People don't want an app.
They want a job done.

Framed around the outcomes people hire MediCare+ for, not the features they tap.

JOB 01

"When I feel unwell at an odd hour, help me know how serious it is — so I can act without panic."

AI triageRisk resultInstant consult
JOB 02

"When a doctor advises a medicine, get it ordered and delivered — so treatment actually starts."

E-prescriptionPharmacyDelivery
JOB 03

"When I manage my family's health, keep everyone's records in one place — so nothing slips."

ProfilesRecordsReminders
JOB 04

"As a doctor between patients, give me context, notes and prescribing in one flow — so I save minutes."

Pre-visit briefDiagnosisRx builder
09 — Product Strategy

A north star everyone could row toward.

One vision, three bets, and the principles that kept every decision aligned.

Vision

Make trustworthy healthcare feel as immediate and continuous as messaging a friend.

North Star Metric

Weekly Care Journeys Completed — a symptom resolved within the platform.

Business goals
  • Lift booking conversion & repeat use
  • Raise doctor utilization & retention
  • Open pharmacy + lab revenue lines
Product goals
  • Cut time from symptom to care
  • Build trust into every touchpoint
  • Keep one record across the journey
Success metrics
  • Journey completion rate
  • D30 / D90 retention
  • Consult → fulfillment attach rate
10 — Stakeholder Ecosystem

The care graph, mapped & live.

Every actor connects through one record and one orchestration layer — a change in any node updates the whole system.

MEDICARE+ Care Graph one record · one engine Patient+ Family DoctorApp + Portal AdminOps & Compliance PharmacyFulfillment LaboratorySample & results InsuranceClaims EmergencySOS & hospitals
11 — User Journey Mapping

The patient journey — emotion and all.

Mapping action against feeling shows exactly where trust breaks, and where a small design move pays off most.

Symptom
Triage
Find doctor
Consult
Treat
Follow-up
anxious unsure hopeful reassured in control cared for
Action
Notices symptom
Action
Runs AI check
Action
Matched to doctor
Action
Video + notes
Action
Rx → delivery
Action
Reminders + records
Pain
"Is it serious?"
Pain
Trust in AI
Pain
Which doctor?
Pain
Tech friction
Pain
Two more apps
Win
It remembers
Design bet

Invest the most design effort at the two earliest, most anxious moments — triage and doctor trust — to unlock the entire downstream funnel.

12 — Service Blueprint

What the user sees vs. what makes it work.

A single consultation, blueprinted across the line of visibility.

TriageBookingConsultPrescribeFulfill
FrontstageAI chat & risk cardDoctor match & slotVideo roomRx summaryTrack delivery
BackstageTriage model + safety rulesAvailability engineNotes workspaceInteraction checksPharmacy & courier
SystemsNLP + risk scoringScheduling DBWebRTC + EHRDrug DB + e-signInventory + payments
Admin / supportMonitor AI accuracyVerify doctorsEscalate emergenciesCompliance auditRefunds & disputes
13 — Information Architecture

Five tabs. Zero dead ends.

The patient IA is deliberately shallow — every primary job is one tap from home. Depth lives only where clinical detail demands it.

5
Primary tabs
3
Max depth · taps
52
Total screens
0
Dead ends
Patient AppBottom-tab navigation
MediCare+ · Patient
01Home
  • AI Symptom Check
  • Book a Consult
  • Daily Summary
  • Reminders & Tips
02Appointments
  • Upcoming
  • Past
  • Manage / Reschedule
03AI Assistant
  • Chat
  • Lab Insights
  • Report Reader
04Records
  • Health Timeline
  • Documents
  • Family Profiles
05Profile
  • Pharmacy
  • Labs
  • SOS / Emergency
  • Settings
Admin ConsoleSidebar navigation · 8 modules
01Dashboard
  • KPIs
  • Alerts
  • Activity
02Doctors
  • Verification
  • Management
  • Profiles
03Patients
  • Directory
  • Details
04Consultations
  • Live Monitoring
  • Appointments
05AI Monitoring
  • Model Accuracy
  • Health Checks
06Pharmacy · Labs
  • Inventory
  • Orders
07Payments
  • Commissions
  • Refunds
  • Subscriptions
08Analytics
  • Healthcare
  • Geo
  • Compliance
14 — User Flows · Live Data

Watch the data move.

The core loop, animated. A symptom enters the AI engine, branches by risk, flows through a doctor to fulfillment — then feeds back into one record that makes the next visit smarter.

Routing · LOW RISK Patientsymptom in AI ENGINE Triage NLP · risk score Low riskself-care + reminder High riskSOS · emergency Doctorconsult · diagnoseprescribe MEDIUM Fulfillmentpharmacy · labsdelivery Health Recordone source of truth
Tap a flow to trace its steps
15 — AI Integration · The Differentiator

AI as the front door —
not a gimmick bolted on.

Six capabilities turn an anxious "is this serious?" into a guided, safe next step — and feed structured context to every downstream product.

AI · 01

Symptom Checker

Conversational intake mapping symptoms to concerns.

AI · 02

Triage & Risk

Low / medium / high routing with safe escalation.

AI · 03

Health Assistant

Always-on chat for questions and prep.

AI · 04

Lab Result Reads

Plain-language interpretation of values & ranges.

AI · 05

Report Analysis

Upload a report; get a summary & flags.

AI · 06

Medication Guidance

Dosage, interactions and adherence nudges.

Triage architecture · input → processing → output
INPUTS

Symptoms · history · vitals · profile

→
PROCESSING

NLP · risk scoring · safety rules · confidence gate

→
OUTPUTS

Risk level · path · doctor brief · specialty

→
BENEFIT

Faster, calmer, safer decisions

Symptom intake
Adaptive follow-ups
Risk result & routing
Lab interpretation
Built safety-first: when confidence is low or red flags appear, the AI steps back and hands off to a human — never the reverse.
16 — UX Decision Making

Every choice had a "because."

Senior design is defensible design. A sample of the calls — and the evidence behind them.

Why AI first?

74% wanted severity before booking. Triage de-risks the scariest moment and pre-qualifies demand.

Why 5 tabs?

Hick's Law — cap primary choices at the five jobs people actually do; push rare actions deeper.

Why bundle fulfillment?

⅔ abandoned when meds lived elsewhere. Rx → delivery in-app protects outcome and revenue.

Why show verification?

61% distrusted online doctors. Badges and ratings surface early to convert trust into bookings.

Why a dedicated SOS?

Emergencies can't compete with everyday UI. A separate, high-contrast path removes ambiguity.

Why one record?

Continuity drove 3.2× retention. Doctors get context; patients never repeat themselves.

17 — UX Laws Applied

Principles, applied with intent.

Where established UX laws shaped real screens — named, and used on purpose.

LawHow we applied itWhereImpact
Hick's Law5 tabs; progressive disclosure of rare actionsGlobal navFaster starts
Fitts's LawThumb-zone CTAs; 48px+ targetsBooking, SOSFewer mis-taps
Jakob's LawFamiliar chat, cart & video patternsAI, pharmacyLow learning curve
Miller's LawHealth summary chunked into 3–4 cardsHome, recordsBetter recall
Doherty ThresholdSub-400ms feedback; AI "thinking" statesTriage, searchPerceived speed
Tesler's LawSystem absorbs complexity (routing, checks)Rx, triageLower load
Peak–End RuleReassuring result + clean summaryRisk, summaryHigher CSAT
Aesthetic–UsabilityCalm clinical visual system builds credibilityWhole productTrust → conversion
18 — Accessibility

Healthcare must work for everyone.

Built to WCAG 2.1 AA — the people who need care most often face the steepest access barriers.

AA
WCAG 2.1 target
4.8:1
Min text contrast
0px
Min touch target
0%
Critical flows labeled

Accessibility scorecard

Color contrast98
Touch targets96
Screen reader94
Dynamic type92
Keyboard nav95

Inclusive design checklist

  • Color never the sole carrier of meaning
  • Dynamic Type scales to 200% without truncation
  • Full VoiceOver / TalkBack on critical flows
  • Confirm destructive & emergency actions
  • Keyboard-navigable web portals
  • Plain-language medical copy
19 — Wireframes → Final Design

From grey boxes to a product people trust.

Fidelity climbed only as questions got answered — structure first, then hierarchy, then the calm clinical skin.

01 · Low-fi — structure
02 · Mid-fi — hierarchy
03 · Hi-fi — final UI
20 — Design System

"Pulse" — one system, every surface.

Token-driven across iOS, Android and web — a button means the same thing for a patient, a doctor, or an admin.

Color tokens
Primary / 500
Primary / deep
Info / blue
Ink / bg
Surface
Warn
Critical
Color = product palette · Type · Space Grotesk + Inter
Display / 34
Heading / 22
Body / 16 · Inter — calm, legible, high-contrast for clinical reading.
LABEL / MONO 12 · UPPER
Spacing

4-pt base · 4/8/12/16/24/32/48

Grid

Mobile 4-col · Web 12-col · 24px gutter

Components

68 components, fully variant-mapped

States

7 states per interactive token

Component states · Button
Default Hover Focus Disabled Secondary
21 — Final Solution

The product, end to end.

Four connected surfaces, one record. Swipe each rail to walk the real screens.

01 · Patient App — iPhone 16 Pro

Onboarding
Onboarding
Home
AI Assistant
Doctors
Doctor profile
Select slot
Confirmed
Video consult
Summary
Pharmacy
Checkout
Reminders
Records
Family
Emergency
SOS

02 · Doctor App — iOS · mobile

Splash
Splash
Login
Login
OTP verify
OTP verify
Registration
Registration
License upload
License upload
Verification
Verification
Approved
Approved
Dashboard
Dashboard
Appointments
Appointments
Patient profile
Patient profile
Video consult
Video consult
Diagnosis
Diagnosis
Prescription
Prescription
Earnings
Earnings

03 · Doctor Portal — Web console

Dashboard
Dashboard
Schedule
Schedule
Appointments
Appointments
Patient directory
Patient directory
Consultation workspace
Consultation workspace
Prescription builder
Prescription builder
Revenue
Revenue

04 · Admin Console — Web

System overview & KPIs
System overview & KPIs
Doctor verification
Doctor verification
Doctor management
Doctor management
Patient management
Patient management
Consultation monitoring
Consultation monitoring
Emergency monitoring
Emergency monitoring
Pharmacy management
Pharmacy management
Payments & commissions
Payments & commissions
22 — Edge Cases

Designed for the bad day, not just the demo.

Healthcare can't degrade into a blank screen. Every failure state has a safe, human next step.

Doctor no-show

Auto-rebook, priority queue, goodwill credit.

Network failure

Offline banner, retained inputs, audio-first fallback.

Payment failure

Slot held 10 min, clear retry, alt methods.

Emergency detected

Red-flag interrupts flow → SOS + nearest hospital.

No doctor available

Notify-me, alt specialties, async message.

Verification rejected

Specific reasons, re-upload, support — never a dead end.

Appointment cancelled

Instant refund rules, one-tap rebook.

Medication missed

Escalating reminders, caregiver alert, refill nudge.

23 — Validation & Usability Testing

We tested, we learned, we shipped better.

Rounds with real patients and doctors — what broke, what we fixed, what shipped.

0%
Task success rate
▲ from 71%
0
SUS score
"excellent" band
−0%
Time triage → consult
vs. first prototype
0
Moderated sessions
2 rounds

Task completion by flow

AI triage96%
Book appointment93%
Order medicine90%
Doctor onboarding89%

Findings → iterations

  • Risk result felt clinical → added plain-language reassurance + next step.
  • Verification trust gap → surfaced badges earlier in matching.
  • Pharmacy hand-off confusion → unified Rx → cart into one screen.
  • SOS discoverability → promoted to a persistent, high-contrast entry.
24 — Design Iterations · Before vs After

Iteration is the work.

The same screens, versions apart — every change earned by a test or a constraint.

BeforeAfter

Triage entry

Buried under "Book" → promoted to a hero home action. Triage starts ▲ 2.4×.

BeforeAfter

Doctor cards

Name + price → verified badge, rating, next slot. Booking intent ▲ 31%.

BeforeAfter

Rx → buy

3 screens + handoff → one screen, in-app cart. Attach rate ▲ 27%.

Across two rounds, task success climbed from 71% → 94% — the difference between a prototype and a product.
25 — Product Impact & Business Results

The numbers that mattered.

Outcomes from the pilot cohort. Modeled figures are labeled projected.

▲ Conversion
+0%
Appointment conversion
▲ Completion
0%
Consultation completion
▲ Adherence
+0%
Medication adherence
▲ Retention
3.2×
D90 retention vs. baseline
▲ Productivity
+0%
Doctor consults / day
▲ Revenue · proj.
2.1×
Revenue per active user
Care journeys completed / week
1.0k
M1
1.8k
M2
2.7k
M3
3.6k
M4
4.4k
M5
5.1k
M6

The north star moved because the whole loop moved — triage fed bookings, bookings fed fulfillment, fulfillment fed retention.

Pilot cohort, 6 months. Revenue-per-user and projected figures modeled from conversion, attach-rate and retention deltas; labeled where not directly measured.

26 — Future Roadmap

Where MediCare+ goes next.

What ships once the loop is trusted — and where the AI gets smarter.

PHASE 01 · NOW

Deepen the loop

  • Chronic-care programs & care plans
  • Insurance claim automation
  • Wearables → live vitals in triage
PHASE 02 · NEXT

Scale the network

  • Multi-language & rural expansion
  • Pharmacy & lab partner marketplace
  • Doctor AI co-pilot
PHASE 03 · LATER

Predictive care

  • Proactive risk detection & outreach
  • Population health dashboards
  • Outcome-based incentives
27 — Reflection as Lead Product Designer

What I'd carry into the next build.

The calls I'd repeat, the ones I'd change, and what leading this taught me.

Key learnings
  • In health, trust is a feature — built in the first 30 seconds.
  • Continuity beats novelty: the record is the moat.
  • AI earns its place only when it knows when to step back.
Trade-offs
  • Leaner emergency module to launch safely, sooner.
  • Depth on triage over breadth of specialties at v1.
  • Manual compliance review where automation wasn't yet safe.
Challenges
  • One system for three very different roles.
  • AI helpfulness vs. clinical safety & liability.
  • 180+ screens coherent under one system.
Leadership takeaways
  • Frame every decision with a measurable "because".
  • Bring research, not opinions, to align stakeholders.
  • Own the outcome — research through business impact.
My job wasn't to make it pretty. It was to make a hesitant person at 2am feel safe enough to act — and to make the business work while they did.