Organisations are making people decisions without people data.
Annual surveys are months old by the time they're read. And gut feeling is structurally unverifiable. Neither detects the withdrawal pattern that precedes a resignation — and both suffer from the same flaw: asking too much, too rarely, to get honest answers.
The most consequential failures in an organisation — burnout, attrition, disengagement — are structurally invisible to leadership until they have already occurred.
87% of employees are not fully engaged.
The cost of not knowing.
Burnout has detectable early signatures. Without a continuous signal, the intervention window closes before leadership knows it opened.
Leadership programmes, restructurings, new team leads — without measurement continuity, no people investment can be evaluated.
Structural problems are routinely blamed on individuals. Without causal separation, the intervention lands at the wrong level.
Capacity and resource decisions are made without knowing the psychological state of the units involved.
Ten adaptive questions per employee per week.
Every question exists in multiple validated phrasings. The system tracks which version each employee last saw — the same question never repeats.
Each session targets what the system knows least about for that employee. Stable topics are skipped unless something shifts.
Ten questions, never interrogative, never the same twice. Short enough for honesty — precise enough that every answer moves the estimate.
Answers are anonymous by default — no individual data is ever visible to management.
What makes it different.
Each psychological dimension per employee is maintained as a live probability estimate — not a score. The system never asks a question it already knows the answer to.
People answer ironically, under social pressure, or not quite honestly. The system accounts for this: uncertainty doesn't distort the signal. It becomes part of it, statistically weighted into every estimate.
No output is labelled until the evidence earns it. The system moves through defined confidence stages — from no data to stable — and never skips one.
is live.
Field validation
underway.
Fully deployed. Measurement, inference, and interpretation run end-to-end — with deterministic fallbacks at every layer.
15,705 unique question variants — tested across 14 psychological topics, every possible way a person might respond, and every stage the system moves through. More edge cases covered than most instruments encounter in a decade of deployment.
Early interest spanning occupational psychology, organisational consulting, academic research, and investment.
Why I built this.
I grew up inside teams, all failing the same way.
First in sport, then in organisations and institutions. The same patterns of mismanagement, the same burnout trajectories, the same wasted potential.
Felt by the people inside. Paid for by the organisation around them.
I'm not a psychologist. I see systems — where they leak, where they compound, where they break. Economics, psychology, engineering: for this problem, you need all three.
As AI absorbs repetitive work, the quality of human judgement and leadership becomes the scarcest asset a company holds.
Measuring it is no longer optional.