Make the days between therapy sessions less lonely, and stop every session starting from scratch.
Case study 01HealthTech
UK (Remote)
MyndHaven is an AI mental-wellness app for the days between therapy sessions. It learns your patterns, then briefs your therapist before you walk in. As co-founder and head of design, I lead its product design from the first interview to the design system.
3 min read · 2025 to now
The 30-second version
For busy readersDesigned an AI companion, a pre-session brief for therapists, client assessments and a moderated peer community.
One connected product, shaped by 20+ interviews and usability sessions, with a waitlist across eight countries.
- 20+research interviews and usability sessions
- 8countries on the pre-launch waitlist
- 1Figma design system
Therapy helps. Everything around it is disconnected.
You journal in one app, track your mood in another, then spend the start of each session catching your therapist up. Between sessions, you're on your own.
I own the experience, end to end.
- 01 · research
Ran 20+ interviews and usability sessions with prospective users and therapists.
- 02 · design
Designed check-ins, the journal, AI chat and voice, assessments, therapist booking and community.
- 03 · system
Built the Figma design system, ran design QA and set the privacy rules.
Team: Me and three software engineers
Therapy wasn't the problem. The gaps were.
Three things came up in almost every interview.
- Insight 01
Tools that don't talk
Nothing people logged ever reached their therapist.
- Insight 02
Starting from scratch
Recapping the week ate into every session.
- Insight 03
Warm, but useful
Too soft feels like a void. Too clinical feels like a form.
The calls that shaped the product
A check-in that feels like a conversation
People open the app in vulnerable moments.
Shipped
Gentle wording and carefully ordered moods, because wording changes how people answer.
What the therapist actually sees
The companion learns a lot. How much should it share?
Shipped
Patterns, moods and key moments, with privacy rules set before the screens.
Before, during and after every session
The companion spots patterns, like Sunday-evening spirals, turns them into a brief before each session, and keeps a shared summary after it. Tap any screen to see it full size.
- Conversational check-ins
- Briefs, not data dumps
- Privacy first
From interviews to a design system
→ git log --releases
- 01Listen20+ interviews
- 02MapCompanion, brief, summary
- 03DesignCore flows and community
- 04SystemDesign system and QA
✓ Pre-launch, with a waitlist at myndhaven.com.
A product people are already waiting for
- 20+research interviews and usability sessions
- 8countries on the pre-launch waitlist
- 1Figma design system
Instead of spending the start of a session recapping the week, the therapist already knows it.
What MyndHaven taught me
Designing for someone on their worst day changes everything, down to the order of a mood list.
- What I found interesting
How much a single word in a check-in changes what people are willing to share.
- What was challenging
Making AI output something a therapist can trust and review in seconds.
- A mistake I made
I refined the therapist brief before testing it with therapists, then had to rework its structure.
- What I learned
Put AI output in front of its real reader before polishing how it looks.
Where it's at: Pre-launch, with a waitlist at myndhaven.com.
Next case study
