The Leadership Problem Behind AI

This episode, I explore why school AI progress is being slowed not by tools or teacher capability, but by leadership, culture, and the redesign of work, drawing on Microsoft’s 2026 Work Trend Index.

Transcript

SPEAKER_00 0:00

If this episode makes you think, please let us know in the comments and support us by subscribing and leaving a review. Thank you. Today we are exploring something I've been thinking about after spending the day with the leadership team and board of governors at St. Julian's School in Lisbon, Portugal, and reflecting on Microsoft's 2026 work trend index. The question at the heart of this is simple. If the capability is already there, why are so many schools still struggling to make meaningful progress with AI? I think many schools are at this exact point now. Their people are using AI. The confidence is there in pockets, the experimentation is happening, the tools are available, but what comes next? Because the pattern is familiar, isn't it? Licenses get bought, a pilot group gets trained, a flagship case study gets shared, a helpful list of suggested uses lands in the all staff email. It all feels like movement. It all feels like progress. And then six months later, when you really look closely, most of the organization is still doing roughly what it was doing before. That's what I've been noticing more and more. The things holding schools back are not mainly about capability anymore. They're not even mainly about access. They're about leadership. They're about culture. They're about whether the organization has actually decided what better work looks like now that AI exists. What really struck me from Microsoft's 2026 Work Trend Index is that only a relatively small group of AI users sit in what the report calls the frontier zone, where individual skill and organizational readiness reinforce each other. A much larger group sits in the middle, using AI here and there, showing some confidence but not really working inside a system that helps that use turn into transformation. And the finding that really matters, I think, is this one. The report suggests that organizational factors like culture and manager behavior drive about twice the impact of individual capability. That should make every school leader stop for a second. Because in schools, when progress stalls, our instinct is often to assume a training problem. Staff need more CPD, they need more confidence, they need another workshop, they need more examples. But what if that is only the surface of the issue? What if the deeper problem is that people are being asked to work in new ways inside structures that still reward the old way? That's a very different challenge, and it's why I think the old digital adoption playbook isn't working with AI. Traditional digital transformation was often a tool problem. You introduced a new system, trained people how to use it, and expected behavior to shift in fairly contained ways. A different place to store documents, a new button in the workflow, a better way to file expenses. But AI isn't really like that. AI is not just another tool layered onto the existing shape of work. It changes how we think about the work itself. That sounds abstract until you see it in practice. A school can give every teacher access to a powerful AI tool and still see almost no meaningful shift in planning, feedback, communication, curriculum design, or professional learning. Why? Because the real question was never can staff open the tool. The real question is what are they now expected to do differently? Because that tool exists. That's where I think schools keep getting stuck. One of the first mistakes I see is the case study trap. We point to the most innovative teacher, the most adventurous department, the standout team, and we treat that as proof that the strategy is working. But it usually isn't proof of systemic change. It's proof that highly motivated people can do impressive things. Well, of course they can. The problem is that this often inspires the already converted and quietly alienates everyone else. It says, whether we mean it or not, this is for the unusually confident few, not for the ordinary middle of the staff room. And for me, that middle matters most. Not the tiny percentage who are always going to run with this, and not only those who are deeply anxious and need careful support, but the broad middle who are busy, capable, slightly curious, slightly cautious, and waiting to see whether this is real. If your strategy doesn't move them, it probably isn't a strategy yet. The second mistake is the tool rollout itself. Buying everyone a copilot or Gemini license and calling that adoption. I always think that's a bit like buying an exercise bike and assuming fitness has been achieved. Access matters, of course it does. But access is not the same as changed behavior. Giving someone a powerful tool without changing expectations, processes, or team norms often just means a few people accelerate and everyone else carries on as before. This is why I keep coming back to a very simple principle. Start with purpose, not technology. Start with why, not how. If a school begins with the question which AI platform should we buy, it is already slightly off track. The better question is where are the friction points in the work? Where are teachers spending time on repetitive tasks that drain energy but add little value? Where are leaders drowning an admin that keeps them away from the decisions and conversations that matter? Where are support teams doing work that could be accelerated without lowering quality? Then the next question matters even more. Which parts of this work can be supported by AI and which parts must remain deeply, unapologetically human? That line matters because I'm not interested in replacement, I'm interested in enhancement. I want schools to outsource the doing, not the thinking. Let AI help with the draft, the structure, the first pass, the summary, the pattern spotting, but keep the judgment, keep the care, keep the imagination, keep the discernment. Machines can compute, they cannot wonder, they cannot care. A third mistake is the use case list, and I say this as someone who understands exactly why schools like them. They are concrete, they are practical, they reduce the fear. A list of twenty useful prompts feels helpful. Sometimes it is helpful, but the danger is that it tells people where to stop looking. It encourages imitation instead of redesign. So instead of asking what are ten ways teachers can use AI for planning, I think the stronger question is what does excellent planning look like now that AI exists? What should become faster? What should become deeper? What should become more collaborative? What should become more responsive to student need because the routine parts have been lightened? That moves the conversation from tricks to standards, from novelty to quality. And that's exactly where the Microsoft study becomes interesting. The people in the so-called Frontier group are not just using tools more often. According to the report, they are more likely to brainstorm and refine practices together, share what they are learning, talk openly about where they are failing, and discuss quality standards for AI assisted work. Now that's fascinating to me because none of that is really about software, it's about culture. It's about whether the team has permission and expectation to think together about the shape of better work. Imagine what that could look like in a school. A department meeting that does not just showcase a clever prompt, but looks at an AI assisted lesson resource and asks, is this actually better? Why is it better? What professional judgment improved it? What would we reject? What standard are we applying here? That is a completely different level of conversation. A fourth mistake is thinking an AI coach or a single AI lead can carry all of this. I understand why schools do it. It feels sensible, it gives the work a home, it creates a point person, but if one person becomes the AI person in a school, the organization may be closer to plateau than to transformation, because this cannot sit on the edge of leadership. It has to sit at the centre of it. AI touches curriculum, it touches assessment, it touches pastoral care, it touches operations, communications, safeguarding, inclusion, parental trust, staff workload, professional learning. So this is not really an IT project, it is a leadership question, a whole organization leadership question. And I think that means leaders need to ask much better questions than the ones many of us started with. Not what AI tool should we buy? But what does excellent work look like in this team now that AI exists? What are we no longer willing to accept as normal? And what process change makes the old way unavailable? That last question is the one that really matters. Because encouragement is not the same as expectation. Permitting AI use is not the same as building a culture in which people are expected to think carefully about how it improves the work. If leaders model that openly, I think it changes everything. When a school leader shows how they used AI to analyse a survey, refine a communication or reshape a planning document, and then talks openly about what they changed, what they distrusted, what they improved, that sends a powerful signal. It says this is not a side hobby for enthusiasts. This is now part of professional practice. And that takes us into classrooms too, because once you stop asking are we using AI and start asking what is better work, then assessment starts to look different. Writing starts to look different. Homework starts to look different. Student independence starts to look different. If AI can generate a decent answer quickly, then the job of the teacher is not to ban thinking aids and pretend the world has not changed. It is to design learning that demands judgment, depth, perspective, performance, and care. That is why I keep talking about process and performance, not just product. The final piece of work still matters, of course it does. But now we also need to know how students got there, what decisions they made, what they refined, what they rejected, what they can explain live. The real value is not only in what the machine produces, it is in how the learner responds. And I think the same applies to organizations. The real value is not in whether AI appears somewhere in the workflow. It is in whether the people doing the work are thinking more clearly, creating more effectively, collaborating more intelligently, and spending more of their time on the parts of education that only humans can do well. The Microsoft report also points to a tension that I suspect many educators feel right now. Plenty of people know they need to adapt. They sense that standing still is risky, and yet many still feel safer sticking with current goals than redesigning work around AI. I completely understand that. Redesign is harder than adoption, it is slower, it is messier, it asks more of leadership, it forces choices, but I also think that is now the real work. Policy alone is not enough. A policy can tell people what is allowed, it can create boundaries, it can reduce confusion, all useful. But policy is not strategy. Strategy says where we are going. It says what kind of school we are trying to become. It says what better teaching, better learning, better feedback, better leadership, and better use of time should now look like. And perhaps that is the simplest takeaway from all of this. The goal is not more AI use. The goal is better work, better work for teachers so they have more energy for the students in front of them. Better work for leaders so they can focus less on administrative clutter and more on culture, clarity and direction. Better work for students so they are not trained to outsmart machines, but to outthink them. That for me is the real shift. Not tool deployment dressed up as transformation, but the slower, deeper work of redesigning expectations, processes and standards, less obsession with novelty, more attention to quality, less focus on who has the cleverest prompt, more focus on what kind of thinking and learning we now want to protect and grow. The capability is there. In many schools that is no longer the problem. The next step is leadership. The next step is deciding what excellent now means, and then build in the culture, the habits and the expectations that make it real. That's all for today. Thanks for listening.