Is Your Staff Data Ready For AI?
School leaders are betting on AI. Their staff data isn't ready for it.
Ask a head of school any of these questions:
Who is due for an appraisal or observation this term? Who is qualified to teach high school biology? Whose first-aid certification lapses before the winter break? How many staff are not yet compliant with the school's safeguarding policy?
What usually follows is a pause, then "let me check with someone." It isn't that the information doesn't exist. It's scattered across a recruitment portal, three spreadsheets, a shared drive, a filing cabinet, and the memory of whoever has been here longest.
Here is where the AI conversation tends to take a wrong turn. Faced with questions like these, schools are increasingly reaching for AI - one silo at a time. An AI-assisted tool for recruitment here, an AI notetaker for observations there, a smart compliance tracker somewhere else. Each makes its own corner faster and a little more clever. It feels like progress, but it doesn’t move the needle.
A smarter silo is still a silo. Point AI at a single function and you get a more sophisticated version of the fragmentation you already had - just more tools that still can’t answer a question spanning all of them. Look at those questions again: not one is about intelligence within a single function. They require connection across several functions. Appraisal status, subject qualifications, certifications, and safeguarding compliance are four facts about the same people, sitting in four different places. Connecting scattered data points into a systemic picture is exactly what AI is good at. Unfortunately, a siloed approach to AI adoption prevents this connection.
The Hunt For Data
Most HR leads know the feeling of going on a hunt. A board member wants headcount by division. A new director of teaching and learning wants to understand observation history. An accreditation team needs a list of who has completed which training. None of it should be hard. But the hiring record sits in one system, the contract in another, the appraisal notes in a divisional folder that looks different for every section, employee ID numbers in a Google Drive or SharePoint folder, and professional development carefully maintained by one diligent employee.
That last point is important. In many schools, the real staff record is not a system at all - it’s a person. Ask how PD is tracked and the answer is a name, not a process. When that person leaves, the institutional memory leaves with them, and their replacement starts the hunt all over again.
People who know me hear this ad nauseam: People are not systems. Systems support people. Good systems are designed for people.
Heads of school feel a different version of the same problem. They are accountable for every adult in the building, yet few could produce an accurate, current picture of that workforce on demand. Not because they don’t care, but because no single source of truth exists to care about. You can’t lead what you can’t see, and right now most leaders only see fragments.
What Connection Makes Possible
Get the foundation right and the use cases arrive on their own. Identify retention patterns across divisions, not just within one. Flag renewal and compliance deadlines before they become an emergency. Answer "how many of our staff are due for appraisal and have not been observed this year?" in seconds rather than in a day of emails - because that question, which spans two functions, finally has one place to land.
And the reverse is the warning. Point a capable tool at six disconnected spreadsheets and a filing cabinet and it will do exactly what a human does; fail…just faster. You can’t automate a mess; AI will just hand you a quicker, more confident version of that mess. At least it will make you feel great about yourself for asking (That’s sharp framing that shows your strong leadership insight - you’re right to ask those questions).
There is a better reason to get the foundation right. Generative AI tools are probabilistic: they return the most plausible answer, not a guaranteed-correct one. For drafting a policy or spotting a trend, a strong best guess is what you want. But many questions need more than a best-guess;
When does this contract end? When was this teacher last observed? How many of our staff meet local legal requirements? Are we compliant with our accreditation standards? Is every adult in the building cleared under our safeguarding policy?
Each of these demands a deterministic answer - exact, verifiable, auditable. "Probably" is not a word you can put in a compliance report or an accreditation submission. Deterministic outputs require data that is structured and connected. If you point a probabilistic tool at scattered documents and data, it will hand you a confident number that no one can stand behind.
On questions that carry real consequences, that is not where you want to be.
For HR and school leadership - AI is a data question. And it’s one they have to answer together.
The Fix Is Not Another Platform
The temptation is to buy something. But bolting a new tool onto the pile rarely helps; it usually just adds yet another place to look. The required shift is in how you think about staff information. A few principles are worth adopting, whatever tools you eventually use.
Treat the staff record as a lifecycle, not a folder. A person's data should flow from application through hire, onboarding, appraisal, development, and eventually offboarding as one continuous thread - not restart in a new tool at every stage.
Give it a single owner and a single home. Fragmentation continues when everyone assumes someone else holds the main copy. Decide where the source of truth lives, and make that decision explicit and known.
Build it to survive turnover. The test is simple: if the person who currently "just knows" left tomorrow, would the record still stand? If the answer is no, you do not have a system - you have a dependency (not to mention a burdened staff member!). Remember: People aren't systems. Systems support people.
Structure it - and connect it - so it can be questioned. Notes buried in prose cannot be queried, and facts trapped in separate systems cannot be cross-referenced. Consistent fields, consistent categories, and consistent practice across divisions are what turn scattered documents into something you - or a tool - can interrogate across functions.
The Groundwork
None of this is glamorous, and it starts with a decision rather than a purchase: HR and school leadership agreeing that knowing who works at your school, in a structured and durable way, is foundational, not administrative.
Now, full disclosure: I have spent enough time staring at this particular mess that we have gone and built software for it at Skoolspot. In fact, our schools tell us it does the job rather well.
Shameless self-promotion notwithstanding, the principle holds; a staff record has to be a lifecycle, single-homed, turnover-proof, and connected enough to question.
For heads of HR, it is the end of the hunt. For heads of school, it is the difference between managing a workforce and merely hoping about one. And for anyone serious about using AI well, it is the decidedly unglamorous groundwork that has to come first.
The schools that get real value from AI over the next few years will not be the ones with the flashiest tools in each department. They will be the ones that connected the data underneath them. The teams that did the necessary, unglamorous work of getting their data in order, so that when they ask AI the hard questions, there is a real, defensible answer waiting.
Who knows, you just might deserve the sycophantic adulation that accompanies it.
Resources: