A development tool — not a wellbeing benefit
Measure impact. Scale storytelling and reframing skills.
Storylab is a coach your people talk to in private. It asks for one real story in the first minute, gets genuinely interested in the answer, and builds attention, storytelling and reframing over weeks. You get completion and skill movement. You never get a transcript.
Sold as cohorts of 24–40 · SSO via Okta, Entra or Google · POPIA and GDPR aligned
One exchange, two replies
“I went to my dad’s place on Sunday. It was fine. Same as always, he didn’t really ask about the job thing.”
The obvious reply
It sounds like there may be some unspoken tension with your father, and “fine” might be doing a lot of work in that sentence. What would it mean to feel truly seen by him?
What Storylab asks
What room were you in?
Then a chair. Then the thing his father was holding. They arrive at the tension themselves, holding evidence — which is the only version that changes anything.
Running cohorts with
- Meridian Logistics
- Kestrel Retail Group
- Nyala Health
- Brightline Contact
- Verity Foods
Two products, one purchase
The room they sit in is nothing like the room you sit in.
Your people get a quiet, single-column conversation with no dashboard, no modules and no scoreboard. You get the measurement layer. The wall between the two is the product.
What your people open
The Quiet Room
Context remembered from Tuesday
You said you’d get through one conversation without talking about yourself. Did you manage it?
Say it or type it…
No sidebar. No streak counter. No modules to work through. One question at a time, and it waits.
What you open
Admin
Active seats
142
of 200
Reaching stage 4
68%
+9 pts
Rung-3 done
58%
+4 pts
Aggregates only, minimum cohort of eight before any figure renders. No session content reaches this screen, ever.
The Exposure Ladder
Five rungs, escalating on one axis: how much it asks you to do differently.
A course can front-load the hard work, because enrolling was the commitment. A conversation cannot, because leaving costs nothing. So trust accrues first and the ask grows after.
- 1
Notice
Change nothing, look harder
“Photograph three people”
- 2
Record
Capture something in the moment
“Write one line when something catches you”
- 3
Interact
Alter one exchange
“Have a conversation without talking about yourself”
- 4
Act
Do something you wouldn’t
“Say yes to the thing you were going to decline”
- 5
Commit
Put it in the calendar
“Book it”
Three things that make it work on a shop floor.
Voice-native
Speak it, don't type it.
Most of the workforce this is built for will not write four paragraphs in a text box, and shouldn't have to. Sessions run on voice by default, transcribed as you go. Typing stays there for anyone who prefers it.
Trait-aware memory
It remembers what you said, not what it decided about you.
The coach can quote you word for word from eleven weeks ago. It can tell you what you said, when, and in what words — and it will not tell you what that means about you. That boundary is enforced, not encouraged.
Escalating difficulty
Five rungs, and nobody is pushed up them.
Assignments escalate on one axis: how much they ask you to do differently. Notice, record, interact, act, commit. Declining is data, not failure — a cohort that stalls at rung three is telling you something about pace.
How it works in detail — the seven stages, the memory schema, and what the coach is forbidden from saying.
What we measure
Four numbers that predict whether it worked.
Soft-skills tools are usually sold on engagement, because engagement is easy to produce. These four are harder to produce and harder to fake, which is the point of choosing them.
Deliberately not measured
Session length. Daily active use. Message volume. All three reward the wrong behaviour — a good session can be nine turns, and an entry written out of obligation is worse than no entry at all.
- Assignment completion, by rung
- The strongest signal there is. Someone completing rung-three assignments is having different conversations at work. That is the actual claim.
- Unprompted specificity growth
- Concrete details per hundred words, measured at the first session and again at the sixth stage. The practice working, measured directly.
- Self-correction rate
- How often someone catches something they had missed. It climbs when the thing is working, and flatlines when it isn't.
- Return after a 14-day gap
- The honest retention question. Not whether they opened it yesterday — whether they came back after they stopped.
The wall
Nobody at their company can read a session. Including you.
Half the workforce assumes a tool like this has a reporting line to HR, and they are right to check. The product only works if that assumption is false, so it is false by architecture rather than by policy.
- Session content is never exposed to any admin role — there is no screen, export or API that returns it.
- Every admin figure is an aggregate over a minimum of eight people, suppressed below that threshold.
- Skill movement comes from a pre/post self-report instrument, not from reading what anyone wrote.
- A learner can delete their vault, and the aggregate recomputes without them.
Start with one cohort of forty.
Twelve weeks, a named cohort lead, and a completion report at the end you can take to a budget review. If completion comes in under fifty percent, we would rather tell you than renew you.