Synetic — Case Study — Jihan Raihana

Why a designer–engineer couple started building healthcare software

I freelance as a brand designer for a clinic in Indonesia — three doctors, two services (acupuncture & aesthetic). I've been with them since they were a logo on a napkin, through their grand opening.

Working inside the clinic gave me something rare: a daily, unfiltered view of how a small healthcare business actually runs.

The owner kept hitting the same wall:

→A generic clinic-management tool — fine for cashier and attendance, but not built for medical workflows.
→Patient records lived in Excel and Word documents.
→Stock tracking was manual and error-prone.
→Real EMR software was out of budget.

This is the reality for most small clinics in Indonesia. Existing EMRs are priced for hospitals; small clinics fall through the cracks.

After two earlier attempts that never reached real users — Lilist, a finance app, and JAKA, a POS system — my husband and I made a different choice this time: build for a clinic we already have access to. Real users. Real workflow. Real constraints. That's how Synetic started.

Designing for four roles, three specialty contexts

DOCTOR: Consultations, medical records, specialty documentation, treatment plans, prescriptions
RECEPTIONIST: Patient booking, check-in, payment
ADMIN: Master data, financial reports, consumable stock (BHP), clinic dashboards
THERAPIST: No dedicated workspace yet — tracked for commission when assisting treatment

The harder design problem was specialty workspaces. An acupuncturist documents radically different things than an aesthetic doctor — yet both are doctors in the same clinic.

ACUPUNCTURE· SOAP notes· Acupoint documentation· Needle count & placement· Multi-session plans

AESTHETIC· SOAP notes· Before / after photos· Visual treatment plan· Consent forms

GENERAL (FUTURE)· SOAP notes· Prescription· For commercialization

Instead of building three separate apps, I designed one platform with swappable workspace modules — and split it by how much is shared:

SHARED — 60%auth · patient records · billing · calendar

SPECIALTY — 20%workspace layer

ADMIN — 20%reporting

But the structure didn't land in one shot. My first draft in Figma Make was specialty-first: each workspace had its own dedicated flow — acupuncture went Complaint → Diagnosis → Point Mapping → Needles → Plan.

When I walked the clinic team through it, they kept reaching for the same word: SOAP. Indonesian doctors use the SOAP framework — Subjective, Objective, Assessment, Plan — across every specialty. My specialty-first structure was technically novel, but practically a stranger to them.

I restructured on the spot. SOAP became the universal backbone; specialty content (acupoint mapping, before/after photos, clinical findings) became plugins injected at the Assessment and Plan steps.

DESIGN PRINCIPLE

SOAP as foundation. Specialty as plugin. Never the reverse.

Adding a new specialty no longer means redesigning the workspace — just designing its injection plugin for Assessment and Plan.

PLATE 02 — ACUPUNCTURE SOAP WORKSPACE

synetic.app/workspace/acupuncture

SHIPPED — SOAP BACKBONE + ICD-10 (M54.5) + ACUPUNCTURE INJECTIONFIG. 02

Shown in part — Synetic is a live product, so the full workspace stays in the clinic.

PLATE 03 — BREADTH: MULTI-ROLE MODULESFIG. 03

Receptionist — registering a new patientMODAL

synetic.app/patients

Admin — clinic finance & reportingBUILT IN CODE, NO FIGMA

synetic.app/admin/finance

What it means to design with AI as a co-designer

Six years of graphic design taught me to be a perfectionist about Auto Layout — every frame, every component, every spacing token, pristine. Then I tried to ship a healthcare app with a team of two before the pilot clinic opened.

We first leaned on Figma Make to win the work: it turns a design into a clickable prototype instantly — perfect for walking the clinic owner through the main flow and features, and closing the deal. The trouble came after the yes: building production straight from that Figma Make output wasn't viable for two people on a budget.

The cost hit twice. The generated code was messy enough that refactoring it into production with Claude burned through tokens — and Figma Make's own subscription added up on top. For a two-person team, that math simply didn't work.

So we found a leaner path — and it's now how we actually ship. I redesign every screen frame-by-frame in Pencil, with Claude as my co-designer; I push to GitHub; and my husband builds straight from it with Claude. The whole pipeline runs on a single Claude subscription — Pencil itself is free. The entire Synetic MVP 1 my husband is building was designed this way, in about a month.

THE HANDOFF LOOP

01Design in Pencil· me — frame by frame, with Claude

↓

02Push to GitHub

↓

03My husband builds with Claude

↓

04Production deploy

↓

05Pilot clinic feedback· real doctors & receptionist

⟲ feedback loops back into Pencil · one Claude subscription, Pencil is free

This isn't the workflow I'd have chosen 12 months ago — it took letting go of tools I was precious about and rebuilding how we work. But it's the only reason a two-person team can ship a healthcare product at all.

A design decision I'm proud of — and had to let go of

THE DESIGN I ALMOST SHIPPED

A doctor selects a diagnosis from ICD-10. The system auto-traces the corresponding acupuncture points on a 3D body diagram. The doctor confirms or adjusts. Treatment is logged.

It would have been beautiful. It would have removed an entire category of cognitive load from the doctor's day. It would also have been a six-month engineering project on its own.

“This is the right idea. We can't ship it in MVP.”— MY HUSBAND, ON THE SPEC

KEPTICD-10 integration — high value, low costStructured text input: points needled, needle count, session notes

CUTThe auto-traced 3D body diagram — six months we didn't have

The doctor still gets structured documentation. We still get to ship.

WHAT DESIGN SCHOOL DOESN'T TEACH

90% of being a useful designer in a small team is knowing what to cut.

PLATE 04 — THE AMBITIOUS VERSION I CUT

synetic.app/workspace/body-mapping

› My Patients ›Consultation Room

Medical Diagnosis (ICD-10)Diagnosis Code

Search by code or name (e.g., M54.5 or Low Back Pain)

Common ICD-10 Codes for Acupuncture

M54.5Low Back PainM54.2Cervicalgia (Neck Pain)M79.1Myalgia (Muscle Pain)M79.3Panniculitis (Trigger Points)M25.5Joint PainM75.0Adhesive Capsulitis of Shoulder (Frozen Shoulder)M77.9Enthesopathy (Tendinitis)M19.90Osteoarthritis

Body Mapping VisualizationView Mapping History

Pin LegendCritical Pain (8–10/10)Moderate Pain (5–7/10)Mild Pain (3–4/10)Nerve TargetTrigger PointFollow-up PointClick anywhere on the body to add a pin. Click the colour dot to change type. Toggle to show previous-visit pins.

AMBITIOUS — auto-trace ICD-10 → body points · CUT for MVPFIG. 04

↓ what shipped instead

PLATE 05 — THE SHIPPED FIX: ICD-10 INSIDE SOAP

The idea survived the cut — ICD-10 lives in the Assessment step, and the doctor selects acupuncture points by hand in Plan. Structured documentation, without the six-month build.

DONE✓Doctor workspace — Acupuncture (V2 SOAP, shipped)✓Receptionist module (booking, check-in, payment)✓Admin module (built directly in code)✓Auth & multi-role architecture✓Prototype with realistic data, ready for testing

IN PROGRESS▣Doctor workspace — Aesthetic (V1 designed; V2 deferred)▣Therapist commission logic▣Patient portal (clinic ops first)▣Staff room (payroll, performance)

NOT YET STARTED☐Therapist workspace☐Synetic's own brand identity☐Marketing landing page☐Commercial business model☐New-clinic onboarding

MVP 1 is now in early testing at the pilot clinic — the doctors and receptionist using it day to day. Next is MVP 2, shaped by what we learn in the field.

01Real users sharpen designs faster than any framework. An hour with the receptionist beats a week of personas.
02Healthcare design is mostly invisible work — the most important screen is a patient record loading in under two seconds during a 15-minute appointment.
03AI tools don't replace design judgment. They compress the distance between idea and prototype. What to build, cut, and test is still entirely human.
04Constraints made me a better designer than ambition ever did. The version we're shipping isn't the one I dreamed of — it's the one that exists.