90-Day AI Governance Roadmap for Schools

Short answer

A 90-day AI governance roadmap for schools

A 90-day AI governance roadmap has three phases. In days 1 to 30, name owners, list every AI tool already in use and share short interim guidance. In days 31 to 60, draft the policy, review the tools on your list and train staff. In days 61 to 90, explain the policy to families, review any pilots and hold your first formal governance review.

Use this page as part of a connected review: understand the purpose, consider the people affected, check the relevant safeguards, and follow the related routes for practical next steps.

Setting up AI governance can feel endless, because there is always another tool or question. A fixed 90-day roadmap gives the work a beginning and an end. It is not about launching AI for students or staff. It is about putting the owners, records and routines in place so that every later AI decision has somewhere sensible to go.

Why governance needs its own 90-day plan.

AI use may have started at your school before anyone decided how to govern it. Teachers found tools, familiar software added AI features, and questions from families began arriving. Setting up governance at this point means catching up while the school keeps running. Without a time limit, the work tends to stretch on, with drafts that never quite finish and meetings that keep revisiting the same questions.

A 90-day roadmap fixes that. It is long enough to do the essential work properly and short enough to hold attention. It also separates two jobs that are easy to blur. Rolling out AI access is about tools and training. Setting up governance is about owners, records and routines. This roadmap covers only the second job, so it can run alongside any rollout or before one begins. A school that already has AI in classrooms can still use it, because the first phase starts by finding out what is already happening.

The most useful step on day one is booking the first formal review for around day 90. Once that meeting is in the calendar, the earlier phases have a single purpose: arriving at that review with a working inventory, a draft policy, recorded decisions and trained staff. Each phase below ends with a short description of what done looks like, so the team can check honestly before moving on.

Days 1 to 30: owners, inventory and interim guidance.

Start with people. The senior leader names a governance lead and the owners for tool review, privacy, classroom use and student support, each with a backup. If the school is forming a committee, hold its first meeting in the first two weeks. Agree how decisions will be recorded and where the shared records will live. Keep this light: a shared folder and a decision log are enough. Tell staff who the owners are by the end of the first week, so questions start going to the right place straight away.

Next, take an inventory of AI already in use. Ask staff, in a no-blame way, which AI tools they use for planning, feedback, communication and admin, and which their students use. Add tools from purchase records, apps connected to the school sign-in and AI features inside existing software. The list will be untidy at first. That is fine. For now, the goal is to see what is there, not to judge it.

Then share interim guidance, ideally within the first three weeks. A one-page note can say which tools staff may continue using for now, that student information should only go into approved tools, that new student-facing tools should wait, and who to ask with questions. Done by day 30 means owners are named, the inventory exists, interim guidance has reached every staff member and the day 90 review is booked.

Days 31 to 60: policy draft, tool review and staff training.

With owners and an inventory in place, the governance lead drafts the policy. Build it from what the inventory showed, so it answers real questions rather than guessed ones. Cover roles, how tools are approved, data and privacy, acceptable use for staff and students, how concerns are raised and when the policy will be reviewed. Ask your privacy reviewer or counsel to check the privacy wording before it goes further. Aim for a draft that a new teacher could read in fifteen minutes, with detail kept in linked documents.

At the same time, work through the tools on the inventory. For each one, the tool review and privacy owners ask the same set of questions and record a decision: approved, approved with conditions, paused while questions are answered, or retired. For example, an elementary school's inventory might show several teachers using the same free quiz tool. Rather than banning it, the school reviews it properly and adds it to the approved list with conditions.

Close this phase with a first staff training session, built around the draft policy and the approved tool list. Practice with real tasks and real questions from the inventory. Invite feedback on the draft, because staff will spot unclear parts quickly. Done by day 60 means a draft policy exists, every listed tool has a recorded decision, and staff have had one training session and a chance to comment.

Days 61 to 90: family updates, pilot review and the first review.

The third phase turns internal work into something the whole community can see. Prepare a family summary of the policy: which tools the school uses, what they are for, how children's information is protected and who to contact. Share it through the usual channels, and explain any changes to students in class in words that suit their age. Make the full draft available for anyone who wants detail.

If any tools were approved as pilots in the previous phase, review them now. Ask the teachers involved what helped, what was confusing and whether the conditions were followed. Record the outcome in the decision log. A pilot review does not need to be long. Its job is to show that conditional approvals are followed up, which builds confidence that the tool review process means something. If a pilot needs more time, say so, set a new review date and record it.

Then hold the first formal review. Look at the inventory, the decision log, feedback on the draft policy and any concerns raised. Agree final changes, approve the policy and set the dates for termly checks and the yearly review. Done by day 90 means families have been told, pilots have been reviewed, the policy is adopted and the next review dates are in the calendar. From here, governance becomes a steady routine rather than a project. Thank the people who did the work, and tell staff briefly what the 90 days achieved.

Fix the end date. Put the first formal review in the calendar on day one, so the whole plan works back from it.

Use interim guidance. Give staff a short holding note early, so they are not left guessing while the full policy is written.

Define done for each phase. Write down what must be true before moving to the next phase, and check it honestly.

Keep scope realistic. Aim for a working first version of governance, not a perfect one. Later reviews will improve it.

Done by day 30:

Governance lead and owners are named.

A first inventory of AI tools exists.

Interim guidance has reached all staff.

The day 90 review date is booked.

Done by day 90:

A draft policy has been shared for comment.

Every listed tool has a recorded decision.

Staff have had a first training session.

The first formal review is complete.

Days 1 to 30. Name owners, take an inventory of AI in use and share interim guidance.

Days 31 to 60. Draft the policy, review listed tools and run first staff training.

Days 61 to 90. Communicate with families, review pilots and hold the first formal review.

After day 90. Move to a steady rhythm of termly checks and a yearly review.

Use this roadmap to plan your school's first 90 days of AI governance work and to check, at each stage, whether you are ready to move on.

Common questions

Quick answers to what schools and families ask most about this topic.