AI Consultant for Schools: Role, Cost and How to Brief One

An AI consultant for a school audits where AI could help, drafts the school's AI policy, trains staff, and builds or connects a few office automations. Published general-consultancy prices run from about $3,500 to $8,500 for a fixed-price audit to $1,000 to $1,400 a day for a mid-level strategist. A head should send a one-page brief and compare fixed quotes.

None of those prices comes from an education specialist, so treat them as a way to test a quote rather than a market rate for schools. The bigger lever is the brief: a head who splits the work into named jobs and asks for a fixed price on each gets bids that can be compared line by line. The school figures below are illustrative.

Follow me on Instagram@sagnikteaches

The four jobs a consultant actually does in a school

Bidders often blur these into "AI transformation". Separating them lets you buy only what you need.

Connect on LinkedInSagnik Bhattacharya

1. Audit. A short look at where staff already use AI, which office jobs repeat, what data each touches and what could go wrong. The output is a ranked list and a risk note, not a technology purchase.
2. Policy. A draft AI policy and acceptable-use guidance for staff, covering what may and may not go into tools, who approves new tools and how parents are told. The governors adopt it, not the consultant.
3. Staff training. Sessions and handouts that get staff using approved tools in the same way. This is the job a capable deputy can often do alone, using material such as a staff session plan with tasks and a handout.
4. Automation builds. Setting up absence reporting, letters, enquiry replies or a handbook chatbot, and handing over a run book. The jobs most worth building first are laid out in five office jobs and what each costs.

A school does not need all four at once. A sensible small-school sequence is audit, policy, one build, then training on the tools actually chosen. Training before the policy exists sends staff off with habits nobody has approved.

Subscribe on YouTube@sagnikteaches

What published prices show, and what they do not

These are the figures I could verify on named pages in November 2026. All three sources are general business consultancies or marketplaces, so they say little about schools. Read them as scale, not as a quote.

Main & Machine's page also describes an audit credit: if you sign a sprint within 60 days, the audit fee is credited against it, capped at 25 per cent of the sprint price. That is a pricing device to notice in any bid: ask whether discovery is credited, and what happens if you stop after the audit. Refonte's guide advises a day or hourly rate for exploratory or uncertain scope, a fixed fee where deliverables are well defined, and a small fixed-fee discovery phase before a fixed quote when the project is uncertain. That is the logic a head should borrow.

What none of these pages tells you is what a school should pay. A school often has no budget line for consultancy, little tolerance for open-ended days and a governing body that will ask for a comparison. The detail on small-business pricing generally is in how much an AI consultant costs for a small business.

Fixed price or day rate: a decision for each job

The pattern is: pay by the day to find out what you need, then by the job once you know. If a bidder will only quote a "transformation programme", ask them to split it into the four jobs and price each.

Three sample budgets for a 220-pupil primary

These are worked sums using the sources above, for an imagined school. They are not quotes. Confirm any figure with the bidder.

Lean: policy and training in-house, one build outside. A deputy writes the policy with free guidance and a template, and runs two 45-minute staff sessions. A freelancer builds absence reporting for a fixed fee in the Contra band for a simple automation, say $500 to $1,500, plus $20 a month in tools. First-year cost: roughly $750 to $1,750, plus the deputy's time.

Middle: a short audit and one build. A fixed-price audit of two days, which at a mid-level day rate of $1,000 to $1,400 (Refonte) comes to $2,000 to $2,800. That is below the low end of Main & Machine's $3,500 audit, which includes more than two days, so a quote in this range for a school is plausible but should be checked for what it covers. Add a fixed-fee build of $1,500 to $3,000. First-year cost: roughly $3,500 to $5,800.

Fuller: audit, policy, training and two builds. Audit $2,800, policy draft at two days $2,000 to $2,800, a half-day training session at a trainer rate of $800 to $1,000 a day, so $400 to $500, and two builds at $1,500 to $3,000 each. First-year cost: roughly $8,200 to $12,100. At this size, ask whether the same work could be staged across two years.

Notice how much of the total is the builds. That is why the first question to settle is whether the jobs are worth automating at all, using the minutes-saved test in a one-term AI plan for a small school.

What a good audit hands you, in a sample

An audit is easier to price once you can picture the output. This is an illustrative page from the audit of the imagined school. A bidder who cannot show you something similar from earlier work, with details removed, has probably not done many.

The last row matters. A good audit tells you what not to automate, and says so plainly. The minute counts are the school's own, gathered by the office over a week, so the audit is partly a way of getting that measurement done properly.

Before and after: two ways to word the same scope

Before, from a vague proposal: "We will deliver an AI strategy and roadmap for the school, including stakeholder engagement and a technology-enabled operating model." There is nothing you can accept or reject.

After, once you ask for deliverables: "We will deliver, by [date], a ranked list of up to ten office jobs with minutes per week and a risk note for each; a 3-page staff AI policy draft with one revision round; and a one-page run book for each automation we build." You can now check each item exists, and you can compare it with another bid line by line.

If a bidder resists this rewrite, you have learned something about what the original wording was for.

Three imagined bids, side by side

Here is how the scoring sheet works on three made-up quotes for the middle budget. The figures are illustrative.

Bid A wins on the sheet despite the data-handling gap, and the right next step is not to sign but to ask A the data question and rescore. Bid C is not wrong in kind, as a day rate suits uncertain work, but without a cap it is the one a governing body would find hardest to approve. Ask for a cap, or a fixed discovery phase followed by a fixed quote.

Try a two-week trial in-house before paying

A school can learn a lot before hiring. Ask the office manager to log one week of repeated tasks, with minutes and the data involved. Have the deputy read the free-tools guidance and try two approved tools on non-sensitive tasks for a fortnight. Then see what is left that nobody has the hours or confidence to do. That residual list is what you brief a consultant on, and it is shorter and cheaper than the list you would have written on day one. It also means the consultant's first conversation is with a school that already knows its own numbers.

The first call: an agenda that keeps it short

1. Five minutes. The school describes the brief and the budget ceiling.
2. Ten minutes. The bidder describes two similar jobs they have done, what went wrong in one, and what they would do differently.
3. Ten minutes. Data: what they would store, where, and who signs what.
4. Five minutes. Handover: who runs it afterwards, and how long that takes weekly.
5. Five minutes. Price structure and any commissions from tools.

Thirty-five minutes per bidder is enough. If one cannot fit the conversation into that time, picture the project.

Costs that are not on the quote

• Tools and usage. Main & Machine's cost guide gives a planning band of $50 to $500 a month for typical small-business tools and model usage, excluding the build itself. For a small school the low end is realistic, since many automations run on free tiers or about $20 to $30 a month.
• Staff time. A consultant needs access to the office manager and the data-protection lead. Budget four to eight hours of their time for an audit.
• Governor and legal review. Policy adoption goes through governors, and a data-protection review of any tool that touches pupil data may need your adviser.
• Maintenance. A build that breaks over a holiday and has no owner is a cost. Ask for a run book and a named school owner.
• Change of mind. A stopped project, or one that follows an audit you did not act on, still costs the audit fee. Check any audit credit and its deadline.

The one-page brief, filled in

A head who sends a single page gets comparable quotes, because every bidder answers the same questions. Here is a completed example for the imagined school, in a form you can copy.

SCHOOL AI SUPPORT: BRIEF FOR BIDDERS
Who we are: a 220-pupil primary, two office staff, 14 teaching staff.
What we want: (1) a half-day audit of office jobs that could use AI; (2) a draft AI policy for staff; (3) up to two small automations: absence reporting and parent letters.
What we do not want: any tool that stores pupil names, medical or safeguarding information; any product we cannot cancel with 30 days' notice.
Data rule: pupil data stays in our existing systems. Bidders must say what each tool stores and where.
Budget: up to $[amount] in total, fixed price per job please.
Timeline: audit by [date]; policy to governors on [date]; builds live before [date].
Our people: office manager (day-to-day owner), deputy head (policy), data-protection lead (sign-off).
We will judge bids on: clarity of deliverables, what the school owns at the end, references, total cost including tools.
Please include: a day-by-day plan, what we must provide, a run book for each build, and what ongoing costs would be.
Reply by: [date] to [office address].

If you would like a template with more fields, the longer version is in writing the one-page brief first. The line "what we do not want" is the most useful sentence on the page, because it tells a bidder what to leave out of their proposal, and tells you something when they ignore it.

How to compare the bids

Score each bid out of five on these six rows, then add them up. The scores matter less than the conversation that produces them.

A bid with a low price and a 1 for data handling is usually the most expensive bid in the end.

Reference questions that reveal what a proposal hides

Ask each referee, by phone, for ten minutes. These questions are better than "were you happy?".

1. What did you ask for, and what did you get? How different were they?
2. Did the final cost match the quote? If not, why?
3. Who runs the thing now, and how much of their time does it take each week?
4. What broke in the first three months, and how did the consultant respond?
5. Did they ever tell you a job was not worth automating?
6. What would you do differently if you were hiring again?

Question 5 is the most revealing. A consultant who has never advised against a build is selling builds.

Red flags in a proposal

• A fee quoted for "AI transformation" with no list of deliverables.
• Pupil names or sensitive information in the sample workflows.
• No mention of the data-protection lead, governors or safeguarding.
• A lock-in term for tools, or a retainer that starts before anything is built.
• Accuracy or savings promises with no baseline: "cuts admin time by half" means nothing without a count of current minutes.
• Reluctance to name a referee.
• Pressure to sign this week.

Checking that the money bought something

Agree measures before the work starts. For the audit, the output is a ranked list of jobs with minutes per week and a risk note for each. For the policy, it is a document governors adopt, plus a one-page staff summary. For a build, it is the same three numbers as in the office-jobs tutorial: calls at the busiest hour, minutes per letter, and questions that no longer reach the office. Measure them again six weeks after the handover. If nothing has moved, ask the consultant to find out why before final payment, which is easier if you held back a final milestone, perhaps 20 to 30 per cent, until the run book is tested.

A last point on independence. Ask whether the bidder earns commission or referral fees from any tool they recommend. Some affiliate programmes are public, and a consultant who earns from a recommendation should say so. If you would like a second reader for your brief and the quotes you receive, that is a use for my AI implementation consultation.

Questions heads ask before hiring an AI consultant

Should a school pay a fixed price or a day rate?

Use a day rate or capped hourly rate when the scope is unclear, such as a first look at where AI could help. Use a fixed fee once the deliverables are written down, such as a policy draft or one automation. A short fixed-price discovery phase first, then a fixed quote, protects a small budget.

Do published consultant prices apply to schools?

They are a rough guide only. The bands come from general business consultancies, not education specialists, and some are Western-market day rates. Use them to test whether a quote is in the right region of cost, then ask the bidder to explain any gap.

Can we do this ourselves with a staff champion?

For a small school, often yes for the policy and staff training, using free guidance and a willing deputy. Pay for outside help where the build touches pupil data or several systems, or where nobody has the hours. Be honest about the hours, not just the skills.

What should the school still own after the consultant leaves?

The decisions, the policy, the list of approved tools, the data-protection sign-off and a one-page run book for each automation. A consultant can draft all of them, but accountability sits with the head and governors, and the team must be able to run what was built.

Further reads

• Do I Need an AI Consultant? 8 Signs It's Time to Get Help — Eight signs that outside help is worth paying for.
• What an AI Consultant Can't Do for You, and What You Must Own — What the school must keep ownership of.
• AI Consultant vs In-House Hire: Which Costs a Small Team Less? — Consultant or staff hire, compared on cost.
• Is Gemini Free for Schools? What Is Included and What Costs Extra — What a school Workspace account already includes.
• Free AI Tools for Schools: Who Qualifies and What Data They Keep — Free tools to try before buying help.
• What Does an AI Implementation Consultant Actually Do? — The general version of this role.
• AI Chatbot for a School Website: Parent FAQs From Your Handbook — Turn a school handbook into a parent FAQ chatbot: clean the source, pick a free tier, set the rules, hand over to the office, test.
• Can AI Draft a Headteacher Report to Governors From Your Data? — Turn anonymised attendance, progress and finance tables into a governors' report draft, with a figure-by-figure accuracy check before it goes out.
• Can AI Write School Attendance Letters? A Three-Stage Sequence — A three-stage attendance letter ladder written by AI with placeholders only, merged from the register by your office, and signed off by the head.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Main & Machine's pricing page and AI consultant cost guide; Refonte Learning's AI consulting pricing guide; Contra's freelancer hiring page for automation work. Figures are from general business consultancies, not education specialists.