Where does your school actually stand on AI? Why leaders need a baseline before a strategy

Governors, boards, parents and inspectors all want to know what your school is doing about AI. Most school leaders can describe their intentions. Far fewer can say, with evidence, where their school stands today: which AI tools are in use, who approved them, what data they touch and whether practice matches policy. That gap between intention and evidence is the pain point we hear most often, and it is the right place to start.

Leaders are being asked for an AI strategy before they know their starting point

The pressure on school leaders has shifted. Two years ago the question was whether to engage with AI at all. Now the expectation is a plan: a strategy, a policy, a training programme, a position for parents.

The difficulty is that a strategy is a route from one place to another, and many schools have not yet fixed the first point. Leaders are writing plans for the school they assume they have, rather than the one that actually exists. When that happens, the strategy looks complete on paper but does not survive contact with how staff and pupils are really using AI.

This is not a failure of effort or care. It is a sequencing problem. An honest baseline has to come before the strategy, not after it.

The AI already in your school is harder to see than you think

Ask most leadership teams how many AI tools are in use across the school and the answer tends to be a short list of the obvious ones. In practice, AI arrives through several routes at once:

• Existing platforms adding AI features. Tools the school approved years ago may now include generative AI that nobody formally reviewed.
• Individual staff adoption. Teachers sign up for tools with personal or free accounts because they save time, often with good intentions and no route to approval.
• Pupil use outside school systems. Pupils use AI at home and on personal devices, and that use shows up in homework, coursework and wellbeing conversations.
• Department-level decisions. A department buys or trials a tool and the wider school never hears about it.

None of these routes is unusual, and we have written before about the coming AI churn in EdTech and staff building their own AI apps. The leadership problem is what happens when they combine: the school's real AI footprint becomes invisible to the people accountable for it.

Confidence is not the same as evidence

Many leaders feel broadly in control of AI. The school has discussed it, there may be a working group, and there are no obvious incidents. That sense of confidence is understandable, but it is not something a governor, inspector, accreditor or parent can check.

The test is whether you can answer questions like these from a record rather than from memory:

• Which applications in use across the school include AI features?
• Which of those process personal data or children's data?
• Who approved each one, and on what basis?
• Has a Data Protection Impact Assessment (DPIA) been completed where one is needed?
• Which vendors use school data to train their own models?
• Which staff have had AI training, and which roles have not?
• What would we tell a parent who asked which AI tools their child uses?

If several of these answers depend on asking around, the school has confidence without evidence. That is the position that leaves leaders exposed when something goes wrong, or simply when someone asks.

Without a baseline, every AI decision is made in isolation

The cost of a missing baseline is not only reputational. It quietly undermines the decisions leaders are already making.

• Policy is written for imagined use. An AI policy drafted without knowing current practice tends to be either too permissive or too restrictive, and staff work around it either way.
• Training is aimed at the wrong people. Without knowing who is using what, professional development defaults to a general session for everyone rather than targeted support for the roles carrying the most risk.
• Vendor review stays reactive. Tools are reviewed when someone raises a concern, not because the school knows which products carry the highest privacy, safeguarding or AI risk.
• Progress cannot be shown. If there is no starting point, there is no way to demonstrate to governors or boards that the school is in a better position this year than last.

Each decision may be sensible on its own. Without a shared picture of where the school stands, they rarely add up to a coherent approach.

What a useful AI baseline covers

A baseline does not need to be a large audit. It needs to be honest, shared across the leadership team and recorded somewhere it can be revisited. In our work with schools, a useful baseline looks at five areas:

1. Use. Which AI tools and AI-enabled features are in use, by whom and for what purpose, including tools adopted informally.
2. Responsibility. Who owns AI decisions today, and whether leadership, safeguarding, IT, privacy and teaching are all represented in them.
3. Policy and practice. What the school's policy says, and how closely day-to-day practice in classrooms and offices follows it.
4. Vendors and data. Which products handle personal or children's data, what the vendor does with it, and whether that has been assessed at product level.
5. Capability. How confident staff, leaders and governors are in understanding AI risk, and where the gaps are by role.

The output should be short enough for a leadership team to discuss in one meeting, and specific enough to show which two or three actions matter most this term.

What schools should be doing now

• Put the baseline on the leadership agenda before the strategy. Agree as a team that the first job is to establish where the school stands, not to finalise a plan.
• Ask, do not assume. Run a short staff survey or department check-in on which AI tools are used and why. Make clear the purpose is visibility, not discipline.
• Start an inventory of applications, including AI features in existing tools. Even a basic list is more useful than none.
• Identify the highest-risk products first. Focus vendor review on tools that process children's data or are used directly by pupils.
• Compare policy with practice. Note where the two diverge and treat that as information, not as a breach.
• Map responsibility. Check that safeguarding, privacy, IT and teaching leads all have a voice in AI decisions, not only one person or department.
• Record it and set a review date. A baseline is only useful if it can be revisited and compared against next term.

Navigating the AI Frontier Series 2: starting with an honest assessment

This is the thinking behind Navigating the AI Frontier Series 2, our six-session live webinar series for school leaders running from 14 October to 25 November 2026.

Part 1 focuses on governance and leadership. Session 1, Leading AI Together, asks how well schools are currently governing AI and how to assess current practice. Sessions 2 and 3 move on to policy, practice and people, and to managing AI vendors with confidence. Part 2, co-hosted by Dan Fitzpatrick, author of The AI Classroom, turns to the classroom: protecting learning, rethinking assessment and building a shared schoolwide approach.

Each session runs at times for UK and Europe, and for North America. You register once for all six, and recordings are shared with registrants who cannot attend live.

How 9ine helps

The 9ine platform is designed to turn an AI baseline into a working record rather than a one-off exercise.

• Application gives the school a single inventory of approved EdTech, including which tools use AI, with a request workflow so new tools are reviewed before use rather than discovered after.
• Vendor provides product-level assessments covering privacy, AI use and model training, safeguarding and cyber security, so leaders know which tools carry the most risk.
• Governance includes a pre-built AI readiness project, turning baseline findings into assigned, trackable actions and building the defensibility leaders need when decisions are questioned.
• Academy LMS provides role-based AI training pathways, so capability gaps identified in the baseline can be addressed and completion evidenced.

Series registrants can also request a Guest Pass to the 9ine platform, which includes the AI Readiness Toolkit and AI Strategy Template, to start building their own baseline alongside the sessions.

Register for Navigating the AI Frontier Series 2

For further guides, templates and insights, explore our resources library. If you would like to discuss your school's approach with our team, book a meeting.