Fifteen years at the frontier

Fifteen years ago, I built a spreadsheet.

It was September 2011, and I was watching schools lose whole weeks to the mechanics of their own data, chasing a single number across seven tabs. The work that mattered, understanding how their children were really doing, was buried under the work that did not.

So I built a tool to take some of that pain away. That was Tracker+. A rough Excel file, but the idea underneath it has not changed in fifteen years: the data should serve the people, not the other way round.

There was something else driving me too, closer to stubbornness than strategy. Everyone seemed to have quietly agreed on what a tracker was allowed to be, a wall of colour-coded grids and a Venn diagram or two, and that was that. I refused to accept it was the best we could do for schools. It was not. Not by a long shot. That refusal, more than anything, is what has carried us for fifteen years.

Because tracking was never really the point. The point was always the hard questions, the ones a head or a teacher wants to ask but rarely can, because reaching the answer means reaching data that is too scattered, too slow, or too painful to get to. Those questions go unasked, not because people do not care, but because the toil in the way is too great. And that is the word I keep coming back to. Toil. The weeks of input and analysis were never the work. They were the tax on it, quietly accepted as simply the way things are.

Eighteen months ago I predicted that AI would compress weeks of this work into hours, and it felt like ambition with a long road ahead of it. This summer, my inbox told me otherwise. One trust CEO wrote to say the data work was done in record time, earlier than ever before, and that the heads had found it effortless. Data that took thirty minutes to enter now takes five. Analysis that took a week now takes an afternoon.

The toil is collapsing. It is not a prediction any more. It is happening.

Here is what I am proudest of. When we set out to bring AI into a pupil tracker, there was no map for how you do this responsibly, or well. So we worked from first principles and invented it as we went. We built a way to pseudonymise every child's data before a single word of it is reasoned over, so the thinking happens without the names. No one required that of us. We did it because it was right. We dreamed up Discover, a thinking partner you can ask anything about your school. We taught Tracker+ to draft a headteacher's report, and to write a report home to a parent, translate it, verify its own translation, and weave your words back through it. We put quiet analyses into the pages where you need them, and built a coach that helps teachers write better termly notes. None of it existed before we made it.

There is a lazy word for that, and the word is novelty. I do not accept it. A novelty wears off; it has a time limit built in. What we are doing does not. It is not a novelty, it is the frontier, and the frontier is not a place you reach once and plant a flag. It moves, and we will move with it. I am determined that the pace of our own development keeps step with the pace that frontier AI itself now sets. That is a promise about who we intend to be, not a description of what we have built.

And it is for everyone in your school, from the head reading the whole picture to the class teacher who just wants a straight answer about their own class. One thinking partner, for all of them, that never tires and is never too busy, and that will tell you honestly when the data cannot bear the weight of the question, because your hardest questions deserve a real answer, not a confident guess.

Fifteen years ago, some of you trusted a rough spreadsheet made by someone who could not stand watching good people waste their time. Many of you are still here. Thank you. Truly.

The toil is collapsing. What is left, when it goes, is the thinking. And this is the part that excites me: a year ago, half of what Tracker+ does today would have sounded like a wish. The ceiling I refused to accept has gone, and what comes next is limited by one thing only, our imagination.

After fifteen years, it feels like we are just getting started.