Confidential, early-stage
2025
B2C
Healthcare
Narrowing an ambitious health roadmap into an investor-ready MVP
A health app that helps people understand their own health data and prepare for doctor visits. As founding designer, I worked with the founder and two part-time developers to get from a roadmap with five feature areas to an MVP we could build and put in front of investors.
Product name withheld. Designs shared with the founder’s permission.
How it started
I joined an early-stage health app as its founding designer, working with the founder and two part-time developers. The founder wanted one product to replace the tools people juggle to manage their health, like fitness apps, patient portals and medication reminders, and his roadmap covered five feature areas:
Real-time health tracking through Apple Health and wearables
Doctor visit preparation
Personalised AI coaching
A portal for sharing documents with doctors
Medication tracking
I owned design and scope. I decided with the founder what went into the MVP, worked through feasibility with the developers, and made the case for each cut.
Problem
Every feature on the roadmap had a reason behind it
The founder had done his own research, and each feature on the roadmap was backed by it. That made prioritising difficult, because none of the cuts were obvious, so I needed evidence of my own.
With no budget for user research, I started with a competitive analysis of tools like Doctolib and Kaia Health. Each one covered part of the problem, and none connected the data across them. That pointed to two likely problems for the people using them:
Their health data was spread across apps that didn't talk to each other
Nothing helped them understand what their numbers meant, so the only next step was a doctor
Constraints
Scoping for a small, part-time team
The team was small, part-time and had no research budget, so each constraint turned into a decision about scope:
One data connection. Keeping five integrations working was a lot for two part-time developers, and one failure would make the whole app feel broken. The product only worked if people trusted its data, so I pushed to build the MVP on Apple Health only and used that to make the case to the founder for fewer features overall.
Testing I could run myself. Without a research budget, I ran two testing rounds with five people each: one midway through the design work and one on the final MVP.
Nothing that reads as medical advice. Anything the app said about someone's data had to be accurate and clearly framed. That set the limits for the AI coach, which I cover below.
Solution
From the full brief to one clear flow
I pushed for a smaller scope from the start, but cuts were hard to agree on while every feature was still an idea. So I designed the full experience the brief described and tested it in the first round.
The Coach screen, the app's home, did worst: four of five participants couldn't say what it was for. It opened with a health score and stats, and the coaching sat five sections down, below a challenge, bookings and health checks.
I designed a simplified navigation and showed it to the founder next to the full version, with the test results. We agreed on a smaller MVP: Apple Health tracking, the AI coach and doctor visit prep. Medication tracking, Google Fit and other wearables moved to later versions. The portal for sharing documents with doctors was postponed to V2, because its technical and legal constraints needed more research.
The new navigation has four tabs: Coach, Learning, My Data and My Doc. Learning holds short lessons that explain what sits behind the numbers, like how daily habits affect heart rhythm.
In the MVP, that became one flow: connect Apple Health, get one clear insight from the coach, and take anything that needs a doctor into visit prep.
The founder added chat to the AI coach partway through the project, and Ask Coach is how people reach it. I scoped it to questions about the person's own Apple Health data, like why their resting heart rate went up this week, and to stop where a question needed a doctor.
The coach on home. It opens from one observation about last night's sleep and turns those numbers into plain advice and small actions, like a reminder after lunch.
Ask Coach on every tip and vital. It opens a chat with the AI coach that already knows what the person was looking at, so they can ask their question straight away.
Doctor questions saved to visit prep. When a question needs a doctor, the AI coach offers to add it to the next visit. Visit prep keeps those questions with a short checklist and patterns from the person's own symptom log, then brings it all into one summary before the appointment.
Three decisions shaped the MVP, one for each step of the flow:
Impact
The MVP that went in front of investors
The two developers built the MVP from my designs, and the founder presented it in the investor pitch.
I kept the round 1 tasks for round 2, so the results could be compared directly. The change came from the order of the Coach screen: one observation and one suggested action first, the numbers below. Doctor visit prep drew the strongest reaction but was still the harder task, so it's where I'd focus the next round.
Reflection
