CognateUK's work began with a fairly straightforward question: what capabilities will people need in a world increasingly shaped by AI?
It was a useful question. It took us beyond knowledge and technical skill towards qualities such as judgement, adaptability and the ability to work with uncertainty.
That question still matters. But as the work developed through programme design, conversations with employers and educators, and early activity with young people, another question began to take more of our attention: how do those capabilities actually develop, and how would we know if they had?
The shift became visible in the design itself.
Instead of asking only what students should learn, we found ourselves asking what they would have to work out for themselves, which decisions they would need to explain, and whether we could see how their thinking changed.
We had not found a better answer to our original question. The work was beginning to change our understanding of the problem.
When the starting question becomes less visible
Most projects need some definition of the problem before useful work can begin. Assumptions are made, a structure develops around them, and attention gradually moves towards implementation.
At that point, the original question can become surprisingly easy to stop seeing.
That is not necessarily a weakness. Constantly reopening the problem would make sustained action difficult. Some difficulties really are problems of delivery.
But practice also produces information that was unavailable at the beginning.
In our case, we could continue to name qualities that seem important: judgement, critical thinking, communication, adaptability. What became harder to ignore was that naming them told us relatively little about how they develop.
Knowing that judgement matters does not tell us what kinds of experience allow it to form. Saying that someone needs responsibility does not tell us what appropriate responsibility looks like while they are still learning. A strong finished piece of work does not necessarily show what happened to the thinking behind it.
The original question was still valid. It was no longer sufficient.
Difficulty, or a different problem?
There is a risk on both sides of this.
If a problem is reframed whenever the work becomes difficult, reframing can become a way of avoiding a more uncomfortable conclusion: perhaps the solution did not work.
The opposite can happen too. Once time and effort have been invested in a particular understanding of a problem, new observations can be interpreted mainly through that existing frame. A framework designed to help make sense of the work can gradually become something the work is expected to fit.
For us, one useful question was what had become visible through practice that we could not reasonably have known at the beginning.
The answer was the second question, about how capability develops and how we would see it.
This concerned the framing of the work itself. If we were simply struggling to teach an already defined set of capabilities, the issue would have been delivery. The practical work was changing what we thought the programme needed to create and what we needed to pay attention to.
That distinction is not always easy to make.
AI may make it less visible still. It can help generate analysis and plausible answers very quickly within the question it has been given. More analysis does not necessarily tell us whether the starting question still deserves the same confidence.
Someone still has to notice when experience no longer fits comfortably inside the frame.
What a framework is for
None of this makes frameworks less useful.
They give people enough structure to work together, and they let an idea meet practice.
Their value may also lie in making assumptions visible enough to revisit.
Our earlier human-capability framing helped us begin. We are not treating it as wrong, nor does moving beyond it validate what we are developing now.
The focus of inquiry has changed. Through CognateUK, we are now paying closer attention to the conditions in which capability might develop, including the questions that first surfaced through the design.
Those questions are still being tested in practice.
But the experience has left us with a wider question about how organisations learn from the work they are already doing.
When a project has been running for some time, the usual question is whether it is on track.
Are we still solving the problem we started with, and does that problem still look the same from where we are now?
This article reflects EFEC’s developing practice and thinking through CognateUK, its capability-formation initiative. It is shared to support discussion and does not represent the views of the schools, employers or other organisations with which EFEC or CognateUK works.
Image: Illustrative image (AI-generated)
