The complaint that reads like a solicitor wrote it.

Complaints have always come. What has changed is the drafting.

A parent who would once have written three paragraphs of frustration can now produce six pages with headings, statutory citations and a request for a response within a stated number of working days. The tone is legalistic. The structure looks like something a firm produced. It often is not, and that is the point: the effort required to send it has collapsed, while the effort required to answer it has not.

Three consequences follow, and they are practical rather than philosophical.

• Volume rises. When drafting takes ten minutes instead of an evening, more complaints get sent, and the same complainant sends more of them.
• Each one costs more to answer. Staff and, where they are involved, the school's solicitors spend hours on matters that would previously have taken minutes, because the letter raises eight points instead of one.
• The legal references are not reliable. AI-drafted complaints frequently cite law indiscriminately and sometimes cite it incorrectly. A school that tries to rebut every citation will spend its time on the wrong problem.

None of this makes the underlying concern less valid. Some AI-drafted complaints are entirely well founded, and a school that treats the drafting method as a reason to dismiss the substance will find that position very hard to defend later.

The complaints trend sits inside a wider shift. Parents are more willing to formally challenge schools, better able to do it without a solicitor, and increasingly prepared to go the distance.

The clearest measure is the SEND Tribunal. Appeals rose from around 3,100 in 2014/15 to roughly 25,000 in 2024/25. That is an eightfold increase in a decade, in a jurisdiction where most parents represent themselves. Waiting times from registration to hearing now commonly run to 20 to 50 weeks, and the Government's SEND reform consultation, which closed in May 2026, proposed keeping the tribunal as a backstop and last resort precisely because the current volume is not sustainable.

The relevance to a governing board is straightforward. Parents who are used to representing themselves in a tribunal bring the same expectations to a school's own processes. They expect a documented procedure, they expect it to be followed, and they will notice if it is not.

This is the part that is most often misunderstood, and it is the reason clerking matters here at all.

Academies are required to have a complaints procedure and to publish it. That requirement sits in the Education (Independent School Standards) Regulations 2014. The DfE recommends a broadly three-stage shape: an informal stage, a formal stage usually heard by the headteacher, and an appeal heard by a panel.

Now the important bit. If a parent escalates past the school, the DfE can generally only consider whether the school followed its published procedure, or whether it breached its funding agreement. It does not re-take the decision. It does not decide who was right about the original incident.

Read that again in the context of an AI-drafted complaint. The complainant may have cited eight pieces of legislation, six of them irrelevant. None of that is what gets reviewed. What gets reviewed is whether the school did what its own policy said it would do, within the timescales its own policy set, with a panel its own policy allowed.

That asymmetry is uncomfortable, and it is the single most useful thing a board can understand about complaints.

In practice, procedural failures are rarely dramatic. They are small, ordinary and entirely avoidable.

1. The published policy does not match what the school does. The website version says twenty school days. The school works to a different rhythm. The parent quotes the website.
2. A stage gets skipped. An informal conversation is treated as the formal stage, so when the parent asks for the appeal they are told they have already had it.
3. The panel is not clean. A governor who has already expressed a view, or who knows the family, sits on the panel. DfE guidance is that where there is any reasonable doubt about impartiality the governor should not sit, and that governors with no prior involvement are generally considered suitably impartial.
4. Nobody recorded the reasoning. The panel reached a decision and communicated the outcome, but there is no record of what it considered or why. Twelve weeks later nobody can reconstruct it.
5. The timeline is not evidenced. Dates of receipt, acknowledgement, response and appeal are scattered across three inboxes.

Every item on that list is administrative. Not one requires a legal judgement. All five are what a governance professional is for.

Two audiences, two different problems, and schools tend to write a policy for the first while ignoring the second.

With pupils

A blanket ban is difficult to enforce and tends to move use out of sight, which makes it harder to teach anything useful about it. A workable written position covers four things: where AI may be used, where it may not, what must be declared, and how a suspected breach is investigated. That fourth point is the one schools skip, and it is the one that gets challenged, because an accusation of AI misuse is an allegation against a pupil and needs the same procedural care as any other. Detection tools are not conclusive, and a decision that rests on one will not hold up.

With parents

Say something, publicly and neutrally, before it becomes contentious. A short paragraph in the complaints policy is enough: the school welcomes complaints in whatever form parents find easiest to write, will respond to the substance of what is raised, and works to the timescales in this policy rather than to deadlines set in correspondence. That single paragraph does three jobs. It avoids any suggestion the school penalises parents for using AI, it sets expectations about timescales, and it gives staff something to point at when a letter demands a reply by Friday.

With staff

Be clear about what may be put into a general-purpose AI tool. Pupil names, safeguarding detail, SEND information and anything from a complaint file should not go into a consumer chatbot. This is a data protection point and it is worth writing down before somebody pastes a complaint into one to help draft the reply.

• Schools face rise in AI-generated parent complaints, Tes.
• School complaints procedures: guidance for maintained schools, Department for Education.
• Complaints about schools in England, House of Commons Library research briefing.
• Continued rise in claims in the SENDIST Tribunal, HCR Law.

Boardmet is a governance support provider, not a law firm. Where a complaint carries legal risk, take advice from your solicitors.

Everything on this page is procedural work, and it is the work that gets dropped when a school is busy.

• Specialist panels for complaints panels and exclusion reviews, clerked independently, with an outcome letter within 3 working days.
• Governance recovery where the policy, the records and the practice have drifted apart.
• Interim governance support when there is nobody holding the procedural work at all.
• One letter, three procedures when a complaint arrives blended with a records request and exclusion representations.
• What the board should be seeing about complaints for the oversight side of the same problem.
• Academy trust governance in 2026 for the duties that changed on 1 September.