AI health & performance platform
Arvic
Co-founding an AI-native platform that tells athletes one thing every morning: am I ready to train today, and why.
At a glance
- 50%daily active usage, up from 20%
- 150+athletes in the validation pilot
- 3platforms, one design system
Challenge
Athlete data is scattered, and no tool says what to do today.
Approach
Interviews with coaches and athletes led to three answers, injury risk, recovery and readiness, explained by an AI coach.
Result
Daily active usage from 20% to 50% in a 150+ athlete pilot.
The problem
Athletes generate more performance and health data than ever, but it lives in different devices and services. WHOOP or Oura read the body of one athlete; Kitman Labs, Zone7 or Firstbeat serve the staff of a club. Nobody connects the two.
Most tools answer with more charts. Coaches told us they need fewer, clearer answers they can act on before a session, with the signals a watch can't read: pain, stress, mood.
Market · who serves athletes, who serves teams
Each tool solves one side. Arvic connects athletes and coaches in one predictive system.
Market size projected to 2031, from Arvic's B4i one-pager estimates. Positions are qualitative.
Discovery · interviews on behaviours and workflow, two different users
A typical week today
- Check-inChat messagesAnswers arrive late, or never
- PlanSheets & PDFsRebuilt by hand when someone is tired
- ReviewSeveral apps & exportsMerging it takes hours, so it's rare
- DecideExperience & gut feelingOverload shows up after the injury
What we heard
Behaviour“Checks each athlete one app at a time.”
Frustration“Every device tells a different story.”
Need“See who needs attention before the session.”
In Arvic →A squad overview first, then each athlete in depth
A typical week today
- Wake upThe watch appA score, but no “so what”
- TrainThe coach's planNo idea if today is the day to push
- FeelNobody asksPain and stress go unreported
- AskSearch, forums, friendsGeneric advice, not about them
What we heard
Behaviour“Checks the watch first thing, then ignores the numbers.”
Motivation“Train more without breaking down.”
Need“One answer before the session, in plain words.”
In Arvic →A verdict, a daily journal and a coach to ask
Synthesis of the discovery interviews, guided by eight areas: profile, training, injury prevention, recovery & stress, mental wellbeing, data, digital tools, feedback. Notes paraphrased.
My role
Role
Co-founder · Lead Product Designer
I co-founded Arvic and lead product and design: vision, research with coaches and athletes, every interface from wireframes to the shipped system, and the brand. I work with a CTO on the models, a performance scientist on the protocols and a developer on the apps.
With Arvic I was selected for and completed three months at Bocconi for Innovation, which shaped the pricing, the go-to-market and the pitch.
Worked with
- CTO
- Performance scientist
- Developer
What I owned
- 01Product vision & strategy
- 02User research
- 03UX/UI & brand
- 04Design system
- 05Front-end in React
- 06Pilot design & analysis
How I approached it
Research showed athletes make one decision every morning. The product answers that first, explains it second, and lets the athlete ask why in plain language.
Key decisions
- 01Three answers only: injury risk, recovery, readinessFewer signals built trust faster than dashboards full of metrics.
- 02Mobile for athletes, desktop for coachesA five-second answer for athletes, the squad and its numbers for coaches.
- 03A daily journal next to the wearablePain, stress and mood matter, and no sensor measures them.
- 04An AI coach that shows its workAnswers cite the athlete's numbers and project two weeks ahead.
- 05A simulator before the planTest the plan on the model, not on an injury.
- 06One system across iOS, Android and webA team of four shipping three platforms at once.
12345- 1A week of readiness at a glance
- 2One verdict first, in plain words
- 3Injury risk, explained without jargon
- 4The main driver and why it matters
- 5What to do today, in two lines
12345- 1The overall risk as one readable ring
- 2A verdict and its change since yesterday
- 3The drivers behind it, ranked
- 4Each driver's trend, coloured by severity
- 5The coach turns the risk into today's action
What I built
Two products on one design system. Athletes get a simple mobile app: a verdict, a journal and a coach to ask, with the technical numbers left out. Coaches get the desktop app: the whole squad first, then each athlete in depth with load, HRV, trends and the plan simulator.
Built in React with Supabase, edge functions for the AI coach and Stripe billing.
- Athlete app: verdict, journal, coach
- Coach app: squad overview
- Injury risk per tissue with drivers
- Recovery & health from wearables
- Arvic Coach with charts
- Plan simulator
Type · Syne, one family for display and body
App icon · the A as a bolt
Primary
GradientColor · dark first, green means good
Status · same scale on every score
Components
The result
- 50%daily active usage, up from 20%
- 150+athletes in the validation pilot
- 3platforms, one design system
After the research-led redesign, daily active usage went from 20% to 50% across a pilot of 150+ athletes, proof that a clear daily answer keeps athletes coming back every morning.
After the research-led redesign