Skip to content

What School Systems Are Trained to See

Add your subtitle here

Key Takeaways

  • A completed task proves the work was done—not what the learner can now do.
  • Coverage protects coherence; capability protects purpose.
  • Strong evidence of teaching is not automatically strong evidence of learning.
  • If you value agency, judgement, and ownership, you need evidence that makes them visible.
  • AI makes completed work an even weaker proxy for human capability.
  • Stop asking coverage evidence to answer capability questions it was never designed to answer.


A school can have strong evidence that learning was organised and still have weak evidence that students became more capable.

The curriculum is mapped. The unit is taught. The task is completed. The assessment is submitted. The report is written. From a systems view, the work looks coherent.

But what are students becoming able to do?

Evidence is not all the same. Coverage evidence answers one question: what was taught, assigned, completed, or assessed? Capability evidence answers a different question: what can students now do with greater independence, judgement, and responsibility?

Both matter. Coverage protects coherence. It helps schools make sure important learning is not left to chance. But coverage does not automatically show whether students are developing the capabilities the school values most.

Coverage protects coherence. Capability protects purpose.

I see this often in school improvement conversations. Schools have strong documentation. They can show the plan, the unit, the assessment, and the result. But when the conversation turns to agency, judgement, ownership, or the ability to use knowledge well, the evidence becomes thinner.

That is not a failure. It is information. It shows what the system is currently trained to see.

The OECD Learning Compass reminds us that future-ready learning is not only about content acquisition. Students need to use what they know with judgement and responsibility. That means schools need evidence that reaches beyond delivery and completion.

AI makes this distinction harder to avoid. A completed task may show that work was done. It may not show what the learner became more capable of doing. If AI supported the work somewhere in the process, the final product may reveal even less about the human capability behind it.

Here is the practical move. Choose one familiar task and draw two columns.

Column one: evidence of coverage. Content taught. Tasks completed. Assessments submitted. Results reported.

Column two: evidence of capability. Decisions made. Explanations given. Revisions justified. Questions pursued. Ownership demonstrated.

If column one is full and column two is thin, you have found the work.

The goal is not to abandon coverage. The goal is to stop asking coverage evidence to answer capability questions it was never designed to answer.

Download the Visible Agency Leadership Paper at https://leecrockett.net/visible-agency.