Work04 · Real-time wellness appWellness app
Lull
A calm coach that reads the day, not your screen
A desktop app for knowledge workers that senses rising stress from voice, typing, and calendar load — on-device — and offers a timely nudge to reset. Personal and private by default, with opt-in anonymized trends a team can learn from.
Try the product
Talk to Lull's calm coach
Tell it how your day is going — it suggests a reset the way the real coach would. Not therapy or medical advice.
A working agent, running live — not a mockup
Lull
Calm coach · live demo
Lull
private · on-deviceTODAY'S LOAD
Calm · 34%You're steady right now. Lull watches the signals so you don't have to.
Share anonymized trend with your team
Off — nothing leaves this device
Interactive prototype — click around, it actually works
Strategy deliverable
The journey we mapped before designing anything.
Every stage from first spark to loyal return — what the customer does, feels, and exactly where the experience leaks. The dip in the curve is the moment we built the work around.
Customer journey map · concept
Theo, 35 — getting through the workday
An engineering manager with a stacked calendar and a creeping case of burnout he keeps ignoring.
Customer journey map for Theo, 35 (getting through the workday), across 6 stages. Sentiment dips lowest at the Skepticism stage (the biggest drop-off) and peaks at the The reset stage.
- Awareness: Fried by 3pm, every day. The customer Notices he's wired, short-tempered, and exhausted by mid-afternoon. Pain point: Stress builds invisibly all day until it's a sick day or a resignation. Our move: An app that notices the build-up before he does.
- Skepticism: Wait — is this spyware?. The customer Reads that it senses voice and typing, and almost closes the tab. Pain point: Monitoring tools are creepy and employer-surveillance-coded; trust collapses right here. Our move: On-device by construction, worker-owned data, team visibility only via opt-in anonymized trends.
- Onboarding: Oh — it all stays on my machine. The customer Sees the privacy model stated up front and installs the menu-bar widget. Pain point: Adoption dies if trust isn't earned in the very first minute. Our move: Lead with privacy, and keep the widget quiet and unobtrusive.
- A real spike: Third back-to-back, jaw clenched. The customer Hits a genuinely tense stretch; the load meter climbs. Pain point: The moment of strain is exactly when you can't self-regulate. Our move: One quiet signal that tells strain from focus — and waits for a natural break.
- The reset: Sixty seconds, and the edge is off. The customer At a break, a gentle nudge offers a 60-second breathing reset. Pain point: Most wellness tools nag at the wrong moment and get ignored, then deleted. Our move: The right nudge at the right moment, one proven reset — then it gets out of the way.
- Habit: I trust it — and so does my team. The customer Keeps it running and opts in to share an anonymized weekly trend with his team. Pain point: Individual fixes don't change a stressful team culture. Our move: Opt-in anonymized trends let a team act on the pattern without surveilling anyone.
Scroll the map sideways →
Engagement
Discovery + ongoing build
Timeline
8 weeks to first release
Surfaces
Menu-bar widget, app, opt-in team trends
Agent role
Calm coach + intervention timing
The problem
Burnout builds quietly during the workday, and by the time anyone notices it's a sick day or a resignation. The tools that claim to help are either useless (a meditation app you forget to open) or creepy (an employee-surveillance dashboard). Nobody had built the version a worker would actually trust enough to keep running.
The approach
We started with the ethics, because they are the product. Every signal is processed on-device, the worker owns the data, and nothing reaches a manager unless the person explicitly opts in to anonymized, content-free trends. Trust is the feature — without it, the app is uninstalled in a day.
The behavioral work defined what to sense and when to speak: voice tension in meetings, typing cadence and error bursts, and back-to-back calendar density, fused into one quiet 'load' signal that knows the difference between deep focus and real strain.
The intervention design is the craft. A nudge at the wrong moment is just noise, so Lull waits for a natural break and offers one small thing — a 60-second breathing reset — that genuinely moves the needle, then gets out of the way.
Key moves
01
Private by construction
On-device processing, worker-owned data, and manager visibility only through opt-in anonymized trends. The privacy model is the product, not a settings toggle.
02
One signal from many
Voice, typing, and calendar density fuse into a single 'load' read that distinguishes deep focus from genuine strain.
03
The right nudge, the right moment
Interventions wait for a natural break and offer one small, proven reset — designed so the app earns its place instead of nagging.
Design targets
- signals never leave the machine
- On-device
- anonymized trends, worker-controlled
- Opt-in
- to ease a stress spike
- 60 sec
A self-initiated concept study. Lull is a fictional app; the prototype above is interactive, the coach below is a working demonstration, and it offers no medical or mental-health treatment.
Next study · 05 Multi-sided platform
Cargo →