Kennisbank

People and skills: the dimension that isn't solved with a training

Why this dimension stands apart

Of the seven dimensions in the maturity measurement, people and skills is the most frequently cited reason why an AI pilot fails to grow further. Not because people don't want to, but because skill is something different from knowledge. Someone can know what a language model is and still not know when the output can be trusted. That gap between knowing and being able to apply is precisely what this dimension measures.

The dimension is connected to other fundamental dimensions. Skills develop with difficulty if the organization itself is not set up for learning with AI, and practicing with AI applications works poorly if the IT infrastructure does not offer an environment to safely experiment. Anyone who thinks only in terms of training budget misses that order.

What the five levels mean in practice

At baseline, a small number of people work with AI tools on their own initiative, out of the organization's sight. There is no shared picture of what does and doesn't work, and what someone learns stays with that person. What baseline means in practice is therefore not a lack of interest, but the absence of structure to make use of that interest.

At foundation, there is a first shared basis: a number of people have built up comparable skills, a common vocabulary has emerged, and separate experiences are occasionally shared. What foundation means in practice is the step from individual use to something that starts to resemble a team that understands the same things.

At activation, people apply AI structurally within their own tasks and recognize where it is and isn't suitable. At insight, the center of gravity shifts: people can not only use AI, but also judge whether an outcome is correct and why. At intelligence, that judgment is no longer a separate step, but part of how work is normally done, spread across the organization rather than residing with a few early adopters.

Where the spread in the plot round comes from

With this dimension, the biggest difference between what a CIO scores and what a team leader scores often arises. A CIO sees the trainings that have been organized and counts that as progress. A team leader sees that that training has changed little about how work is actually done, and scores lower. Both observations are correct; they simply measure a different part of the same dimension. Supply and application are not the same thing, and the plot round exposes precisely that gap instead of averaging it away.

What raising a level depends on

Raising a level on people and skills does not depend primarily on budget for courses. It depends on repetition: people who keep applying AI to real work, not to practiced examples. It depends on time that is actually available alongside regular tasks, and not just allocated on paper. And it depends on an environment in which making mistakes while learning is not a reason to fall back on the old way of working.

In addition, the step depends on whether skill is made visible and shared. Someone who learns to properly check an AI outcome only helps the organization move forward if that way of working ends up somewhere that others can take over. Without that mechanism, any progress remains tied to one person, and disappears as soon as that person changes roles.

This immediately makes clear why this dimension is rarely fixed on its own. How people learn whether someone has performed well, and whether that learning is recognized and rewarded, directly touches on how mature performance management is. And skill also develops more slowly if the data people practice with is unreliable, which is why this dimension is rarely separate from how mature data management is.

What this does not answer

This dimension shows whether the people in an organization are ready to work with AI at a level that fits the ambition. It does not show which part of a specific role AI can take over, and which part must remain with a human. That is a different question, one that is answered per task rather than per organization. The work scan from FTE TO AI calculates which part of the work in a role is suitable to hand over to AI, task by task. Where this page describes whether people are ready to guide that handover, the work scan shows exactly what can be handed over — and that only becomes a meaningful question once the receiving side, as described here, is up to standard.