AI is already in your school. The only question is whether it's governed.
Staff are drafting emails with chatbots. Pupils are using them for homework. Somewhere, someone has pasted something sensitive into a free tool. None of this is scandal — it's what happens when a genuinely useful technology arrives faster than policy. The job of leadership isn't to stop it (you can't) but to convert it from an unmanaged risk into a managed capability. Here is the 30-day plan we use.
Week 1 — Find out what's actually happening
Skip the survey nobody answers honestly. Instead, run short, blame-free conversations with a cross-section of staff: what are you already using AI for, and what do you wish it could do? Make amnesty explicit — the goal is a map, not a disciplinary file. You'll typically find three things: more usage than leadership expected, a handful of genuinely smart use cases worth spreading, and one or two practices that need to stop immediately (usually involving pupil data in consumer tools).
Close the week by writing down your school's red lines: the data that must never enter an AI tool (pupil-identifiable information, safeguarding records, HR matters) and the decisions AI must never make alone (assessment judgements, safeguarding decisions, anything reportable).
Week 2 — Publish a policy people can actually follow
Most AI policies fail by being either bans (ignored) or essays (unread). A workable school AI policy fits on one page and answers four questions:
- Which tools are approved, and for what categories of task?
- What data may never be entered into any AI tool?
- What must a human review before it leaves the building — and who signs it off?
- How do staff disclose AI assistance where it matters (reports, communications to parents)?
Anchor it to your existing data protection and safeguarding policies rather than inventing a parallel universe. Then brief it in ten minutes at a staff meeting — policy that isn't briefed doesn't exist.
Week 3 — Run one supervised pilot
Pick a single, low-risk, high-annoyance workflow. Good first candidates: drafting routine parent communications, generating differentiated practice questions, or first-draft admin documents. Bad first candidates: anything touching assessment judgements or pupil data.
Define success before you start — in minutes saved and quality maintained, judged by the staff doing the work. Run it with a small volunteer group, with human review as a non-negotiable step. The pilot's second output, quietly more valuable than the first: your staff learn what AI is actually good at, which inoculates against both hype and fear.
Week 4 — Train, decide, document
Turn the pilot group into your champions. A single practical training session — real school tasks, live, with the policy on screen — moves more staff than a term of memos. Then make the evidence-based decision: extend the pilot, adjust it, or stop it. Write down which, and why. That one-page record is the start of your AI governance trail, and governors will thank you for it.
What this month buys you
By day 30 you have: a truthful map of AI use, a one-page policy staff have actually seen, one pilot with measured results, a trained core of confident users, and a documented decision. That's not an AI strategy — it's something better: the evidence base an AI strategy should be built on.
The schools that struggle with AI over the next few years won't be the ones that moved slowly. They'll be the ones that let a hundred private experiments run ungoverned while waiting for certainty that never came.