How Do You Answer SLT Questions About AI Without a 40-Page Policy?

Someone on SLT has asked the question you've been half-dreading.

Is AI safe for our marketing and admissions work? Are we putting pupil data into something we don't understand? Are we replacing jobs, or chasing a trend?

You know the tools could save hours on social drafts, enquiry replies and open day copy. You also know you can't walk into a governors' meeting with a shrug and a ChatGPT screenshot. They want a clear answer, not a tech demo.

It's quite simple, really. Most of the time they are not asking for a PhD in large language models. They want to know you have thought about safeguarding, human review, and what "good" looks like before anything goes out in the school's name.

What do governors and SLT actually want to know about AI?

They want reassurance in ordinary school language.

Is pupil or parent data going into public tools? Who checks the output before it is published or sent? Does this replace a person, or help a stretched team of one or two? What happens if the tool invents something? How will we explain this to parents if they ask?

Those are fair questions, and they are answerable on one page if you write marketing and admissions rules you can follow on a Tuesday afternoon rather than a university AI strategy.

A one-page AI policy for school marketing is a short, named set of rules for how your team may use AI tools on website copy, social, emails, enquiry drafts and reports. It is not a classroom AI curriculum and it is not a full IT policy. It is the minimum that lets a Head of Marketing or Admissions say "yes, we are using AI as a co-pilot, and here is how we stay in control" without inventing it in the meeting.

Why does a thin policy beat a thick document nobody reads?

Because a 40-page PDF in SharePoint does not change what happens at 4:30pm when someone needs a social caption or a same-day enquiry reply.

Thick policies often stall the work. People wait for "the official version" while still pasting half an enquiry into a free tool, or they avoid AI and lose another evening rewriting open day emails by hand.

A thin policy that names the tools, the red lines, and the human review step is more useful than a polished document nobody opens. Governors tend to respect that. They can see you are not winging it, and you are not pretending AI is magic either.

What questions should you be ready to answer in the room?

Have short answers ready. Not speeches.

1. Is it safe?

We do not put identifiable pupil, parent or staff data into public AI tools. We anonymise or use approved building blocks. Anything that leaves the school is reviewed by a human who works here.

2. Are we replacing jobs?

No. We are using AI to draft repetitive work so a small team can spend time on visits, relationships and the decisions that need judgement. The school remains accountable for every public message.

3. What about accuracy?

AI can invent details. That is why nothing auto-publishes and nothing auto-sends. Draft, check, then send. Same standard you would apply to a junior colleague's first email.

4. What about DfE or sector guidance?

Keep this honest. Point to your school's own data protection and safeguarding frameworks, and to human oversight sitting with the school. Do not invent a DfE circular you have not read. If your DSL or DPO wants a line in the policy, put their review date on the page.

5. What is the return?

Talk in numbers they already care about: time to first enquiry reply, open day materials ready earlier, commercial posts that ask for a booking, fewer evenings rewriting the same caption. It's not techy waffle they need. It's proof the week got lighter.

What belongs on a one-page AI policy for school marketing?

Keep it to one side of A4 if you can. Name the owner. Date it. Review it termly.

Include:

• Purpose. AI helps draft marketing, admissions and commercial content. Humans approve what goes out.

• Allowed tools. List the ones you actually use (for example Claude, ChatGPT, Grok, Canva). If it is not listed, it is not in use until reviewed.

• Never input. Names, addresses, medical or safeguarding notes, unredacted enquiry forms, staff HR detail, anything that identifies a child or family.

• Always review. No auto-posting. No auto-sending enquiry replies. A named person checks tone, facts, fees language and safeguarding before publish or send.

• Brand and truth. Do not invent testimonials, places, results or quotes. Stick to approved school facts.

• Logging and escalation. A simple note of what you used AI for that week, plus who to ask if a prompt feels close to a red line.

If you want a free starting template you can adapt for marketing and admissions (not classroom use), the AI Starter Guide includes a plain policy shape you can make yours this week.

How do you brief SLT without sounding like you are selling them a gadget?

Lead with the job, not the model name.

Explain the problem in their words: slow enquiry replies, a content calendar that slips every open day week, commercial offers that never get a proper push, governor reports that eat a Sunday. Then show the rule: AI drafts, we check, we own the output.

Bring one example: a before-and-after enquiry reply (details removed), or a caption that used to take forty minutes and now takes ten with a proper edit, plus the never-input list. That is usually enough for a first conversation.

If they ask for training so the team uses the same method, point them to a practical session. On Thursday 17 September I am running the AI for School Marketing four-hour online session for marketing, admissions and commercial leads. Small group. Workspace set-up, brand voice, being found when parents ask ChatGPT, and the workflows that give you hours back. Nab your spot if the date still works for you.

What should you change this week?

Write the one-pager. Even a draft on a shared Doc is better than another week of informal experiments with no shared rules.

Agree the never-input list with whoever owns data protection. Pick one live use case (social drafts or enquiry first replies), run it under the new rule for a fortnight, then take response times or hours returned to SLT.

You don't need a strategy away day to start. You need a clear answer when someone asks "are we being careful?" and a habit of human review that matches how your school already protects families.

FAQ

Do independent schools need an AI policy before marketing teams use ChatGPT?

You need agreed rules before anything with school data or the school brand goes through a tool. A short one-page marketing and admissions policy is usually enough to start. Waiting for a perfect whole-school document often means people use tools anyway, without shared red lines.

What should a school AI policy for marketing include?

Purpose, allowed tools, what never goes in, human review before publish or send, brand and truth rules, light logging, and who to escalate to. If it cannot be followed on a busy open day week, it is too long.

How do you explain AI to school governors in plain English?

Say it drafts, humans decide. No identifiable pupil or parent data in public tools. Nothing auto-publishes. The goal is a faster, more consistent week for a small team, not replacing judgement or relationships.

Is AI safe for school admissions enquiry drafts?

It can be, with anonymised or approved building blocks and a human edit every time. Do not paste full enquiry forms into public tools. Do not auto-send. Safeguarding and accuracy sit with the school, not the model.

If SLT is asking the questions and you want a practical afternoon on policy-friendly workflows, brand voice, AI search visibility and the admin that eats your week, join the next AI for School Marketing session on Thursday 17 September, 10:00–14:00 online. Small group. No jargon. Just tools you can use on Monday morning.

And if you want a free, plain start first (including a marketing-and-admissions policy template you can adapt), the AI Starter Guide is ready to download.