The knowledge your school loses every week

Satya Nadella says the future belongs to organisations that own their learning loop. For schools, the whole organisation is the dataset, and almost all of it currently disappears by Friday.

Satya Nadella recently posted a long note titled “A frontier without an ecosystem is not stable.” His argument has a simple core. In the AI era, the organisations that thrive will not be the ones that pick the best model. They will be the ones that build a learning loop on top of a model, where their own knowledge compounds over time and stays theirs.

He draws a sharp distinction between two kinds of capital. Human capital is the knowledge, judgement, relationships and pattern recognition of the people in an organisation. Token capital is the AI capability that organisation builds and owns once that human expertise is captured, structured and reused. The warning underneath it is blunt. Hand everything to a few generic models and you slowly cede your value to them, the way the first wave of globalisation hollowed out whole industries.

For schools, this lands closer to home than it first appears.

The whole school is talking. Almost none of it is kept.

Think about how much professional knowledge a school generates in a single week.

There is a great deal in lessons, but it is everywhere else too. In the cover lesson where a non-specialist finds a clever way to hold a class. In the department meeting where someone explains why a unit keeps tripping pupils up. In the coaching conversation, the CPD twilight, the moderation discussion and the quick post-observation chat over a cup of tea.

Each of these holds real insight about how teaching and learning actually work in this school, with these pupils, this year. Almost all of it evaporates the moment the conversation ends. We rely on memory, goodwill and the occasional minute-taker. A school’s most valuable knowledge stays locked in individual heads and walks out of the door when people move on.

This is the gap Nadella is pointing at. The knowledge already exists. The school simply has no loop to keep it.

From the lesson to the whole organisation

Starlight began with the lesson. A teacher records a whole lesson, uploads the audio and receives a structured, private coaching report within minutes. The transcript does the heavy lifting, surfacing patterns in questioning, explanation, talk and participation that no observer could hold in their head. We have written before about why feedback should follow the class and its context, not just the teacher.

The lesson is only the first surface. The same engine that turns a lesson into insight can, with clear consent and the right governance, turn any professional conversation a school chooses to record into the same kind of developmental insight. A department’s moderation meeting. A round of coaching conversations. A CPD session. A trust-wide briefing. Each becomes another input to the school’s own learning loop, rather than another conversation that disappears.

The aim is not to record everything. It is to let a school decide what is worth keeping, and to keep it in a form the school owns.

Anonymised, aggregated and trusted by default

None of this works without trust, and trust is the whole game. We say it plainly and we build for it. Individual reports are private to the teacher. Leaders never see individual judgements. What leadership sees is anonymised and aggregated: themes across a department, a key stage or a trust, with thresholds set so that no single person can be identified or singled out.

That is the difference between surveillance and intelligence. Surveillance watches individuals. Intelligence reads patterns. A school can learn that questioning in Key Stage 3 leans heavily on recall, or that a particular misconception recurs across a department, without ever putting a name to a number. Leaders gain a far better basis for CPD decisions, and teachers keep the privacy that makes honest reflection possible.

We take the data questions seriously because school leaders should, and because the answers are a feature rather than an afterthought. We have set out our position on data and the wider questions worth asking.

The loop that compounds

Here is where Nadella’s argument becomes concrete for education. As a school records, reflects and acts across more of its professional life, the aggregated picture grows richer and more specific to that school. The questioning patterns, the explanations that land, the misconceptions that recur and the coaching themes that keep surfacing all accumulate into something the school owns and can build on.

That is institutional knowledge in the truest sense. It is not a folder of policies. It is a living, anonymised picture of how this school teaches at its best, and where it can grow. It survives staff turnover. It grounds CPD in reality rather than guesswork. And because it is built on the school’s own practice, no generic model can package it up and hand it to a competitor down the road. The school is not renting its intelligence. It is building it.

I explore the same theme from a leadership angle on my own site, in a piece on why schools need learning systems rather than more AI tools.

Keeping the human at the centre

A loop like this is only worth building if it strengthens the people in the school rather than replacing their judgement. That is the design principle we keep returning to. The AI captures and surfaces. Teachers and leaders interpret, decide and act. Without that human direction, as Nadella puts it, you simply have compute running in circles.

Used this way, AI does not flatten a school into data. It helps a school see itself more clearly, hold on to what it already knows, and get better at the things that matter, lesson after lesson and conversation after conversation.

The schools that thrive in the next few years will not be the ones with the most tools.

They will be the ones that quietly built a loop around their own expertise and kept it.

Your school is already producing the intelligence. The only question is whether it keeps it.

If you would like to see how this works in practice, you can book a short demo.

Spark Insight with Starlight, and turn the expertise your school already holds into knowledge it keeps.

🎥 Subscribe to our channel here: https://www.youtube.com/@Star21-ai

🌐 Read more on our blog: www.coaching.software

💡 Explore the platform: www.starlightmentor.com

🐦 Follow us on X: @star21starlight

The Insight Engine is written by Adam Sturdee, co-founder of Starlight, the UK’s first AI-powered coaching platform, and a senior leader with responsibility for teaching, learning and coaching. This blog is part of a wider mission to support educators through meaningful reflection, not performance metrics. It documents the journey of building Starlight from the ground up, and explores how AI, when shaped with care, can reduce workload, surface insight, and help teachers think more deeply about their practice. Rooted in the belief that growth should be private, professional, and purposeful, The Insight Engine offers ideas and stories that put insight, not judgment, at the centre of development.

🔗 Connect with me on LinkedIn: https://www.linkedin.com/in/adam-sturdee-b0695b35a/