What an AI workforce means for an education institution
A plain explanation of the AI workforce idea: what the SkoolBro assistant does today, what the planned agents would do, and where people stay in charge.
AI in education
Ten real school office tasks sorted into what an AI assistant does today and what AI agents will take on later, with the privacy rules to agree first.
Ask a school office where the day goes and very little of the answer is about teaching. Someone is counting absences for the principal. Someone else is checking which families still owe fees, or calling parents who asked about admission two weeks ago and never heard back. Most of this work is finding information that already exists in the school's records and getting it to the person who needs it.
AI can take on a good part of that today, and more of it later. Some products promise more than they deliver, so it helps to sort the work honestly into what an AI assistant does now and what is still coming, and to agree privacy rules before any AI tool sees a child's record.
Take a fictional group, Oakfield Schools, as an example. It runs three schools with about 2,100 students between them, and a central office of five people handles admissions, fees and reports for all three. Every Monday the group director asks for the same update: last week's attendance at each school, how much of the month's fees has come in, and how many admission applications are waiting for a decision. Putting that together from three sets of records takes one office assistant most of the morning.
Almost all of that morning goes on looking things up and counting. The answers are already in the records. An AI assistant that reads those records and answers questions in plain words can turn the morning into a few minutes, provided the records are up to date.
The first six tasks are questions, and an assistant that reads your own records can answer them today. The last four are pieces of work an AI agent would prepare for a person to approve. In SkoolBro those agents are on our roadmap and not available yet.
| Task | Who usually does it | Where it stands |
|---|---|---|
| 1. Finding students with low attendance this month | Principal, class teachers | Works today |
| 2. Working out how much of this month's fees is unpaid, by grade or campus | Accountant | Works today |
| 3. Counting admission applications at each stage, such as those waiting for a test or an interview | Admissions office | Works today |
| 4. Checking which staff are on leave this week | Principal, office manager | Works today |
| 5. Comparing exam results between sections and subjects | Principal, heads of department | Works today |
| 6. Finding overdue library books and stock that is running low | Librarian, store keeper | Works today |
| 7. Writing a short summary for the principal each morning | Office manager | On the roadmap |
| 8. Following up inquiries for next year's admissions | Admissions office | On the roadmap |
| 9. Drafting fee reminders for families | Accountant | On the roadmap |
| 10. Spotting absence patterns in a class before they become a problem | Class teachers | On the roadmap |
For each of the first six tasks, someone used to open a register or a spreadsheet, filter it and count. With an assistant, the principal types the question the way they would ask a colleague: "Which students in Grade 9 have the lowest attendance this term?" or "How much of this month's fees is still unpaid at the second campus?" Oakfield's Monday update becomes three questions, one per school, and the office assistant's job becomes checking the answers.
The answers come from the school's own records, so they are only as good as the data. If teachers take attendance on paper and the office types it in a week later, the assistant is answering about last week. If your records are still spread across books and spreadsheets, start with a term-long plan to modernize the school office, because AI cannot read a register book in a cupboard.
An assistant like this should be read-only. It looks things up and explains what it found, and it cannot mark attendance, change a fee or send anything to a parent, which makes it safe to hand to a busy principal from the first day.
Tasks 7 to 10 go further than answering a question. A morning summary means choosing what matters out of everything that happened yesterday. Following up an inquiry means writing to a family. A fee reminder involves money and tone, and an absence pattern needs judgment about which children to worry about. An AI agent can prepare this work, with a person at the school approving anything that goes to a parent or affects money. Be careful with any product that offers to message families without that approval step.
Agreeing a fee concession for a family in difficulty, talking to a parent about a child's behaviour, approving payroll and choosing which applicant gets the last place in Grade 1 all have consequences for a family or a member of staff. They should stay with people after agents arrive. AI can put the facts in front of the person deciding, and the decision and the conversation stay with the person at the school whose job it is.
It also helps to remember that some of the largest time savings in a school office come from ordinary software with no AI in it. A parent hearing the same day that their child was marked absent needs no AI, because the register itself triggers the notice. Count those wins too, and use AI for the questions and drafts where it helps most.
School records hold children's names, family phone numbers, health notes and fee details. Before anyone on your staff uses AI with them, agree rules like these and write them down:
Explain the rules to parents as well. Families are more comfortable with AI when the school can say in two sentences what happens to their child's details. Our post on how SkoolBro keeps each school's data safe covers the wider protections around school records.
SkoolBro's AI assistant is live. It answers plain questions about attendance, fees, exams, admissions, staff leave, library books and stock, using only your organization's records. It sees only what the person asking can see, and it cannot change anything. Before anything is sent to Google's Gemini, names, admission numbers, phone numbers and email addresses in your records are replaced with placeholders, and every message is checked again on the way out. The AI assistant section of our home page shows an example.
The agents behind tasks 7 to 10 are on our roadmap: an AI administrator that writes the principal's daily summary, an admissions agent that follows up inquiries, an attendance agent that spots absence patterns by class and a fee collection agent that drafts reminders for your approval. A teacher assistant for lesson plans and quizzes, and an examination assistant for result summaries by class, are on the roadmap as well. Our overview of the AI workforce for education institutions explains how they will work and where a person approves.
It can be, if the tool reads only the records the person asking may see, replaces names and contact details before anything leaves your system, and cannot change records or message families by itself. Ask any provider to explain each of these in writing.
No. AI takes over much of the looking up and counting, and later the first drafts of routine messages. Office staff still check the answers, talk to families and make the decisions.
Plain questions about attendance, fees, exams, admissions, staff leave, library books and stock, such as which students have low attendance this month. It answers from your own records and cannot change anything.