Is AI amplifying your capability – or replacing it?
The Kaivant-I is the first structured diagnostic built specifically around the augmentation-substitution distinction. It measures not just whether you are using AI, but what that integration is doing to your human capabilities over time.
Two patterns that look identical – until they don't.
When you hand off cognitive work to AI, two very different things can happen. If the freed capacity goes into higher-quality thinking, your output improves and your capability grows with it. That is the augmentation pattern. If AI is handling work that used to build your judgment, the output still looks fine, but the capability that depended on that work quietly stops developing. That is the substitution pattern.
Both patterns look identical in the short term. The difference only shows up later, when it is harder to fix.
No existing assessment measures this. Productivity tools, skills frameworks, and performance reviews all track output. None of them ask whether the person producing that output is growing, or quietly eroding beneath what AI is doing for them.
The Kaivant-I surfaces that distinction. Not a test: a calibration that shows you where your AI use is taking your capabilities, and what to do about it.
What the self-test measures – your personal domains.
The self-test looks only at you. It shows how you personally work with AI, and whether your own abilities are growing or quietly fading. It covers two sides of that.
The first side is how you've built your own way of working to get more done with AI. How much of your routine work have you handed to AI? How quickly and how well do you make decisions? Do you create more coordination work for the people around you, or less? And is your output genuinely improving over time, not just getting bigger?
The second side is the state of your own human capability. Can you still think and reason well on your own, without leaning on AI? How much depth and range do you bring that AI can't replicate? How strong are your working relationships and the trust you build? And could you still perform to a high standard if your AI tools were suddenly taken away?
Underneath both sides sits one more thing: your capacity to keep improving how you work. To notice what isn't working, try something different, and make it stick.
All of this is about you as an individual. It is private, self-assessed, and meant for your own development.
What a Kaivant-I profile looks like.
Fictional scores, real structure. The profile below shows exactly what a Kaivant-I result looks like: two dimension scores, a quadrant position, and a reading of what your pattern means and where to take it next.
Both dimensions are above the midpoint, and the profile position is positive. The proximity to the centre is the developmental signal worth working with: there is real capability here, and real room to build it further.
Stable, no flag in this assessment period. Independent capability baseline is holding. Worth monitoring: the HCD margin above the midpoint is modest, and two consecutive periods of decline would trigger a flag.
A PLA score of 61 says AI is genuinely expanding what you can produce. An HCD score of 53 says your independent capability is above the midpoint, with meaningful room to grow. The Compounding profile at this position is productive: the two dimensions are developing in the right relationship to each other. The question the profile raises is not whether you are doing something wrong, but whether you are being intentional enough about the direction of travel.
These two dimensions have a relationship that matters more than either score on its own. When PLA advances and HCD stays static, the profile gradually drifts toward substitution. Not because anything fails immediately, but because the cognitive work that used to build judgment is increasingly handled by AI, and if nothing deliberate replaces it, the capability that depended on that work quietly erodes over time.
The question to ask is not how much you use AI, but what kind of thinking AI is now doing instead of you. If it is work that was once stretching your judgment, that stretch needs to be found somewhere else. If it is mechanical or administrative work, the freed capacity is available to invest in higher-quality intellectual engagement. The scores cannot tell you which it is. You can.
The modest HCD score at this position is not a concern in isolation. It becomes a concern if PLA continues to advance while HCD stays flat. That divergence is the early signal of the substitution pattern, and catching it at this stage is exactly what the instrument is designed for.
The most useful response to this profile is to identify one domain where you have progressively outsourced the thinking rather than just the execution, and to reintroduce deliberate, human-first engagement with it. Not as a constraint on AI use, but as a deliberate investment in the capability underneath it. HCD advances when judgment is actively exercised and tested, not when it is confirmed by AI output. At this position, that investment compounds quickly.
This profile belongs to you, not your employer. Kaivant-I is a development tool. It is designed to help you understand where your AI use is taking your capabilities, and make deliberate choices about the direction. Individual scores, quadrant positions, and CIR readings are yours alone. The instrument can only produce honest data if you engage honestly with it, which requires knowing that no one else is scoring you on the result.
A development profile – not a performance report.
Your data is private by default. The Kaivant-I produces a development profile for you, not a performance report for your employer. This is not a limitation. It is a design choice on which the instrument's validity depends.
Genuine engagement requires genuine safety
If Capability Independence and Resilience (CIR) results are accessible to employers as performance signals, the instrument generates precisely the behaviour it is designed to detect: people performing their AI-free work for measurement rather than engaging with it genuinely. Privacy is not a policy choice. It is a design requirement for the instrument to produce valid data.
Aggregate patterns only, with individual consent
Organisations offering Kaivant-I as a voluntary development resource can access aggregate, anonymised patterns: enough to understand whether the development investment is working. Individual scores and profiles stay with you.
The signal that overrides the composite.
The Capability Independence and Resilience (CIR) dimension measures one thing: whether your capabilities still function at meaningful quality without AI. Not whether you avoid AI, but whether the capability underneath is holding.
CIR is not designed to penalise AI use. It is designed to detect the substitution pattern: the condition where your AI leverage is growing while your independent capability quietly erodes. Output metrics cannot see this. CIR can.
If CIR declines over two consecutive assessments, your Kaivant-I Score is flagged regardless of how the composite looks. Some conditions are risks, not trade-offs to be averaged out. The flag exists for that reason.
The CIR flag is developmental, not a judgement. It identifies a pattern worth addressing before it becomes a genuine capability deficit that no output metric can see until it is already entrenched.
Declining CIR over two consecutive assessments flags the Kaivant-I Score regardless of composite performance. CIR results are yours alone: a development signal, not a performance metric.
What the self-test can't see – the organisational domains.
The self-test cannot see your organisation, and that is a different picture entirely. No individual can report it on their own. It needs information about how the whole organisation actually runs.
The full assessment measures whether the organisation is genuinely efficient or weighed down by coordination overhead. How fast and how well decisions get made across it. How much of its work runs through AI without constant human hand-holding. And whether its output per person is on an upward path.
It also measures what the organisation is building for the future. How fast it learns, and how well it spreads what it learns. Whether its people are used for real human judgment, rather than rule-following an AI could do. Whether human capability is growing as fast as AI is taking over tasks. And whether people have genuine dignity, a real say, and the safety to challenge the AI systems shaping their work.
It measures one thing more: whether the organisation can actually redesign and rebuild itself when it needs to. That is the difference between an organisation that adapts and one that stalls.
These are structural conditions. They sit above any single person, which is why they can only be measured at the level of the organisation. That is what the full assessment, the Kaivant-O, is for.