The map arrived. The journey did not.

A new episode of The International Classroom goes live with Clare Martin, psychological coach, owner of Goldcrest Training, and author of Wellbeing Ambassadors. The title is "Why workplace wellbeing fails, and how to actually fix it." I went in expecting the usual wellbeing conversation. I got something sharper. Her whole argument is that wellbeing fails for the same reason most school initiatives fail: it gets bolted on. A yoga session. A poster. A fruit bowl. None of it embedded into how the place actually runs.

The moment I keep replaying is a story she told about being booked to embed a culture of wellbeing, and almost nobody came. The session was called resilience and wellbeing, and turning up would signal to managers that they were not resilient enough. The culture was not safe enough to attend the thing designed to help.

I also published piece five of the Better Adults, Better Humans series: the five capacities that actually sit underneath productive work, sustained attention, designed environments, genuine engagement, deliberate practice, and motivation that lasts, and the uncomfortable question of where young people are supposed to learn them. The honest answer is mostly nowhere. They are caught from the adults around them long before they are ever taught, which turns the question back on us: which of the five am I currently modelling the absence of?

The actual document, not the coverage, and I am glad I read the source, because the coverage got it wrong in ways that matter. More than one outlet reported twenty-two competences; at least one invented a fourth domain called "Design AI." Neither is true. The framework has four domains, Engage with AI, Create with AI, Manage AI, and Shape AI, and nineteen competences across them.

What surprised me is the shape of it. The four domains read as a progression, and Shape AI sits at the top, the domain where a young person investigates how a system is built, evaluates it against real criteria, and tries to improve it so it actually serves people. Almost nobody gets there. Most schools, mine included on a normal week, are still living in the foothills of Engage. The framework is quietly ambitious about where children should end up, and quietly silent about how a time-poor teacher is supposed to get them there by Friday.

Three things, all aimed at the same gap between the map and the journey. The first is the AI Literacy Audit, which I have spent this week rebuilding around the new framework. You can now audit your school directly against the real four domains and nineteen competences, and get a scored picture of where you actually sit, with next actions for leaders, teachers, and curriculum teams. A framework tells you the destination. The audit tells you your starting coordinates.

The second is DEEP Track, and this is the one I most want you to look at. It is now open for signups, ready for the new school year, a browser extension and dashboard that lets a teacher capture professional learning in one click while reading or watching it, turn rough notes into credible CPD evidence without inventing impact, and export it straight into appraisal systems like BlueSky without re-keying anything. If you lead CPD in a school, it is built for exactly the staff who never find time to log it.

The third is Lead with AI, still in closed beta, the course for school leaders. It is not a literacy course for teachers. It is for the people who decide whether AI is a side project at the school or a function of the school. If you lead in a school and want a seat, the beta is open, or just reply to this email.

Here is the question I cannot put down. What stops a framework like this from becoming just another spreadsheet column? I have watched it happen. A good idea arrives, schools admire it, and within a term it has been reduced to a mapping exercise. We tag the scheme of work against the domains, tick the competences, file it for the inspection, and nothing in the actual room changes.

This is exactly what Clare was describing about wellbeing, just pointed at a different problem. We fall to the level of our systems. A framework you admire but do not embed is a poster. A framework you build into how you plan, how you appraise, how you talk to each other in the staffroom, that is a system.

And there is a deeper rhyme underneath both. The top of this framework, Shape AI, is not really an AI skill. It is a thinking skill wearing an AI costume. Investigating how something works, judging it against real criteria, knowing what good would even look like. The framework and the coach are asking schools to teach the same underneath skill. Neither of them quite says so out loud.

Before you map anything against the AILit framework, spend twenty minutes with its classroom examples rather than its competence list. The learning scenarios at ailiteracyframework.org are the part the coverage skipped, and they are the part that shows you what Shape AI actually looks like with real children in a real lesson. It is free, it is CC BY, and it is far more useful than any summary of it, including this one.

If you want the wellbeing version of the same idea, Clare pointed me to the free VIA Character strengths survey at viacharacter.org. Evidence-based, takes ten minutes, and a surprisingly good staff-meeting opener.