For product managers · 2-minute read
One PM. Problem to pilot-ready B2B product. Every call written down, with a kill switch.
Sukoon gives Indian HR teams a 2½-minute desk break their people actually take, and weekly proof of use that never names a person. I owned discovery, requirements, safety, build and launch.
The whole case, at a glance
The problemCompanies pay for wellbeing that goes unused, and HR can't prove usage without tracking people.
Who paysHR champions it, Finance signs. Employees use it. Each needs something different, and the product answers all three.
Why it winsCompetes on effort to start (one link, no app, no login) and on a number HR can defend at renewal.
The hard lineNo individual data, ever. Teams under 10 are hidden. That's the product, not a setting.
How I'll knowNorth star: weekly finished breaks per team. Kill if under 30% start by week 2 or under 1 in 5 return in 7 days.
Where it standsLive and pilot-ready. No paying customers yet. HR feedback is the next entry, and this page will show it.
Why this problem: the data says reach and proof, not content
38%of employees use health check-ups even when offered (Plum, 2025, 6,000+ companies)
1 in 3say they lack time for their health (Plum, 2025)
~0measurable wellbeing effect from individual programmes like apps (Oxford, 46,336 workers, 2024)
Insight: wellness fails on reach and proof, not content. So the product competes on effort-to-start and on a number HR can defend, not on more content.
Three people decide. I designed for each.
EmployeeUses it
Needs: under 3 minutes, seated, nothing to install, nobody watching.
HR / TA leadChampions it
Needs: proof of use, zero admin, no privacy risk under DPDP.
Finance / founderApproves it
Needs: cost per person who used it, a clear stop rule.
Every bet has a number that kills it
| Hypothesis | Signal | Kill if |
|---|---|---|
| People take a 2½-minute break if starting costs nothing | Share of seats that start in week 1–2 | Under 30% by week 2 |
| The habit holds past novelty | Return within 7 days | Under 1 in 5 |
| HR will pay for proof without names | Pilot to paid conversion | HR can't name a decision the report changed |
What I measure, and what I refuse to measure
North starWeekly finished breaks per active team
ActivationFirst finish within 48h of launch
RetentionCame back within 7 days
ValueCost per person who started
GuardrailsZero individual data leaves a device · every safety stop handled · teams under 10 never shown
What I said no to, and why that is the strategy
| Not built | Why |
|---|---|
| Manager or individual view | Kills trust, and adoption with it. The whole bet is "never a person". |
| Leaderboards, streaks, badges | Inflate early numbers, then fail; make the data less honest. |
| Dashboard as the main surface | HR doesn't log in weekly. A Friday email is the retention loop. |
| SSO, HRIS sync, roster upload | Each adds an IT approval. One link needs none. |
| AI by default | Typed text would leave the device. Rules-based by default (~900 phrases, four languages); a company can switch on an AI layer such as Claude with its own key and written opt-in. Safety never depends on the model. |
Five versions. Each one changed because evidence said so.
- v0 · Workout companion. Learned people between calls choose by problem ("stiff neck"), not workout type → rebuilt as desk breaks.
- v1 · Safety. My own eval: a keyword list missed 10 of 14 body-pain cases → fixed safety rules the AI can't override, in English and Hinglish.
- v2 · B2B layer. HR needs proof, employees fear tracking → team totals only, teams under 10 hidden, plus the next-smallest.
- v3 · Site. Benchmarked against Apple, Headspace, Calm, Microsoft Viva → plain-language privacy, one primary action.
- v4 · SaaS presentation. Feedback: "it looks artistic, not like a product; I'm not the customer" → removed decorative motion, one product view with three tabs, a demo HR dashboard.
Next entry: results from HR outreach.
Risks I know about, and what covers each
- Privacy vs detail. HR gets less granularity than it may ask for. Accepted: detail without trust gets no usage.
- Novelty decay. Weeks 1–2 inflate. The week-3 check-in and 7-day return metric exist for this.
- No clinical review yet. Disclosed on the safety page; required before any paid pilot.
- Single founder. Pilots capped at six teams.
Next 90 days: ends in a decision, not a roadmap
- Weeks 1–3. 20 HR conversations; one change from the top objection.
- Weeks 4–9. First 6-week pilot, measured against the kill criteria above.
- Week 10+. Kill, pivot, or price. Decided by the numbers, written up here.