Validation Programme · Published Openly · Updated as the Evidence Grows
How KAIVANT earns the right to its claims.
We publish where KAIVANT stands, openly. A measurement standard that earns trust does not claim validity it has not yet demonstrated.
Three claims, and the evidence behind each.
KAIVANT separates what it can show today from what only its own data can show over time. Each claim below names what it rests on. The third stays open until the data settles it, and we say so.
| Claim | What it rests on | Where it stands |
|---|---|---|
| The method works in practice | The managed loop at the core of the method has been used with organisations for many years. The Foundation Paper names its practice source. | Practice-proven. |
| The framework is grounded in research | Published research on expertise, capability and how people work with AI, set out in the Foundation Paper. | Research-grounded. |
| Scores predict outcomes over time | KAIVANT's own data: organisations and individuals assessed, then followed over time. | Being established. Statistical validation grows with every assessment. Not claimed until the data shows it. |
The same applies to both instruments: Kaivant-O for organisations and Kaivant-I for individuals.
Where we are, now.
A framework designed to evolve without drifting.
The KAIVANT framework separates two layers that operate on fundamentally different timescales. This separation allows the framework to respond to a changing world without the intellectual foundation shifting.
The intellectual architecture
Three claims: the two-axis compounding thesis, with its non-compensatory composite logic and the bridge. The augmentation-substitution distinction. The managed loop: compounding lasts only where the organisation runs a deliberate improvement loop around its use of AI. These claims operate at a level of abstraction that sits above specific AI capabilities and is expected to hold for ten to twenty years. Changes require a formal revision process with documented rationale.
Specific metrics and thresholds
The specific metrics, thresholds, benchmarks, and measurement instruments within each dimension. These have a useful life of three to five years and are formally versioned from launch. Annual review is triggered by shifts in AI capability, accumulation of dataset evidence, or the scheduled annual review cadence. Each version release includes a change log, rationale, and transition protocol for organisations in longitudinal tracking.
Help build the evidence base.
Organisations. Assessment data from participating organisations contributes to the validation dataset. Each assessment generates data that improves calibration norms. Each organisation followed over time adds evidence for statistical validation. Organisations that take part in this first phase are recognised as founding validation partners.
The KAIVANT assessments are live on kaivantscore.com. Practitioners: apply to the practitioner cohort on the practitioners page. Organisations: contact us directly.
Practitioners. Founding cohort practitioners are the primary vehicle through which assessment data is generated in this first phase. Each practitioner deployment contributes to the evidence base and informs the indicator review process. Practitioners have direct input to the indicator review at each annual cycle.
The founding practitioner cohort is open now. Applying is free, with no commitment.