Your team is upskilling in AI. You are not confident the right capability is developing.
The training budget is spent. Completion rates look good. Something about whether your people are genuinely better at their work — not just faster at producing it — is still unresolved. That gap is measurable, and it is a different thing to measure than skills training tracks.
The training programme ran. People completed it. AI adoption is up, and by the metrics you can see, the investment looks justified. And yet something in how the team approaches genuinely hard problems, the ones that require working through uncertainty, synthesising without a template, defending a position from first principles, feels different. Not worse in a way you can point to. Different in a way you have noticed.
The standard interpretation is that skills training takes time to show up in performance data. That may be true. But there is another possibility: that the training measured the wrong thing. Learning to use AI tools well is not the same as developing the judgment that remains valuable as AI use deepens. One is a workflow question. The other is a capability question, and capability is not what most AI training programmes are designed to build or track.
The concern arrives before there is data to support it, because the data you have access to does not reach what you are watching. Completion rates and adoption figures measure participation. They say nothing about what is happening to the human capability underneath the tool use.
The Kaivant-I measures two axes simultaneously. Personal Leverage Architecture covers how effectively AI is integrated into each person's workflow: whether the leverage is genuine, whether it is saving time on real work, whether it is compounding. Human Capital Depth covers what is happening to the capability underneath: whether judgment independence is stable, whether people can operate well when the tools are unavailable, whether cognitive range is expanding or contracting.
The distinction that matters in this situation is between augmentation and substitution. In an augmented pattern, AI absorbs execution work and the person invests the freed capacity in harder, more judgment-intensive challenges. The capability develops alongside the tool use. In a substitution pattern, AI absorbs not just execution but the developmental work itself: the synthesis that required thinking, the problems that required working through ambiguity. The output looks similar or better. The judgment underneath is quietly atrophying.
Training completion does not tell you which pattern has formed. The Kaivant-I does. It identifies, at the individual level, whether the patterns of AI use are building the capability the organisation depends on or gradually replacing the work that would have developed it.
Each team member completes the Kaivant-I individually. Results are private by design: the individual sees their full picture first. What you receive at the team level is the pattern across dimensions — which areas of AI leverage are genuinely building, where the substitution signal is appearing, and whether the concern you are carrying is confirmed or contradicted by actual data.
The facilitated session turns that pattern into targeted action. The team reads its results together, identifies the specific dimensions where the gap is forming, and agrees on structural changes: not working less with AI, but restructuring how it is used to reintroduce the developmental work that the current pattern has removed. The interventions at this stage are specific, modest, and measurable against a re-assessment in ninety days.
The goal was not AI literacy. The goal was a team that is more capable because of AI, not one that looks more productive while quietly becoming more dependent on it. The Kaivant-I tells you which of those two things is actually happening.