Every week someone offers to put artificial intelligence into a nursery group. I build software for a living, so I understand the appeal. But I think the conversation is pointed the wrong way. The most powerful intelligence in your organisation is already there, already paid for, and in most groups it is barely being used.

It lives in your nursery managers, your deputies, your room leaders and your practitioners. They know, better than any report, where their nursery stands. They know which room is carrying the others. They know which new apprentice is going to be good and which key person is struggling with her children's records. They know which parent is about to give notice, and why.

The question for a group leader is not how to get more intelligence into the building. It is why so little of what is already there reaches you in a form you can act on.

Three ways it gets locked up

It gets written for you, not for the truth. Sooner or later every nursery manager is asked to write down how good her nursery is. Ofsted expects a self-evaluation. Head office asks for one. Most managers will write the version they think the group wants to read. Not through dishonesty. Through experience. They have learned what happens to a manager who puts her hand up. So you get a document that is accurate in tone and useless in content, and the thing you most needed to know is the thing it was written to hide.

It gets counted instead of asked. Groups are good at collecting what is easy to count. Occupancy, ratios, staff turnover, complaints, accident forms. These matter, and I would never run a business without them. But they are the shadow of the nursery, not the nursery. The number tells you occupancy in the baby room dropped. It does not tell you that the room leader who held that room together left in March and nobody has been brave enough to say so. The person who could tell you that was never asked in a way that made it safe to answer.

It gets filed. After the self-evaluation comes the plan: the list of things the nursery is going to improve, who will do them and by when. Every nursery has one. It gets written after an inspection, a quality visit or a new framework, usually by the manager alone, usually in a hurry. Then it goes in a drawer until the next visit. A plan nobody opens is not a plan. It is a record of what someone once intended, and the knowledge that went into it goes quiet with it.

What unleashing it actually means

It does not mean more monitoring. Monitoring is how you find out what people are doing. Intelligence is what they understand about why. You cannot inspect your way to the second one.

It means four things, and none of them is technology.

Start with the honest mark. Give your managers a simple scale and the same areas and criteria Ofsted's early years toolkit uses, and ask them where they think their nursery sits. Then ask the only question that matters: what makes you say that? Judgement first, evidence second. The order matters. If you ask for evidence first you get a file. If you ask for a judgement first you get a person's thinking, and the evidence becomes the conversation about whether she is right.

Make the framework shared, not the answers. The temptation at group level is to standardise everything. Resist it. What you need in common is the scale and the criteria, so one nursery's judgement can be set beside another's. What you need to vary is the judgement itself. A group where every nursery rates itself the same is a group where nobody is telling you anything.

Ask what is helping and what is hindering, not what is wrong. This is an old idea, and an underrated one. The psychologist Kurt Lewin set it out in 1951: any situation is held where it is by forces pushing it forward and forces holding it back. When you ask a manager what is wrong with her nursery, you get a list of things she cannot change. When you ask what is helping and what is hindering, side by side, you get a picture of forces, and most of the time the fastest improvement comes from strengthening a help rather than fighting a hindrance. That is a different conversation, and people lean into it.

Let the group picture build from the people, not the people from the picture. Most group reporting is built top-down. Someone decides what the board needs to see and the nurseries are asked to feed it. Turn it round. If every manager's honest evaluation is kept live rather than filed, the group view is simply all of them side by side. Nobody is chased for a return. Nothing is re-keyed. And what you see is not a report about your nurseries. It is your nurseries, thinking out loud, in a form you can compare.

Where the artificial kind fits

I am not against it. I use it every day. But its proper place in a nursery group is behind the human intelligence, not in front of it. It can read every report card Ofsted has published for day nurseries since the new framework began and tell you what "Strong" looks like in practice. It can draft the action plan once the manager has done the thinking. It can spot that three of your nurseries have the same thing holding them back and you might want to fix it once.

What it cannot do is know your nurseries. The moment you let a system make the judgement, your managers stop making it, and the thing you were trying to unleash goes quiet again.

The leader's job, restated

If this is right, the group leader's job changes shape. You are less an inspector of your nurseries and more the person who makes it safe and worthwhile for the intelligence already in them to surface. That means receiving an honest "we are not there yet" as the most valuable thing a manager can give you. It means asking "what makes you say that?" more often than "send me the figures". And it means treating every nursery's improvement plan as something you open on a Tuesday because something moved, not something you file in September.

The intelligence is there. It always was. The work is to stop locking it up.

If you want to talk it through

I founded iAbacus and I still spend time with the people who use it, because that is where the thinking comes from. If you lead a day nursery group and any of this rings true, I would be glad to spend an hour with you and one of your managers, working through an honest picture of that nursery. You will leave with the picture whether or not you ever use our software.

If I am not the right person on the day, my colleague Isla Cave is. Isla works with day nurseries and nursery groups full time. She knows the Ofsted early years toolkit inside out, and she is just as good at the part Ofsted never sees: tailoring the criteria to your own group, so the same honest conversation covers occupancy, staffing, parent relationships and whatever else you run the business on. Either of us is a reply away: dan@opeus.org or isla@opeus.org.

Daniel O'Brien founded iAbacus in 2010 with John Pearce, a former headteacher and inspector, and has been Director of Opeus since 1999. The method is used by schools, trusts, colleges and nursery groups across the UK.