AI Marking Software for School Moderation Policies

Moderation is one of the few things in a secondary school that lives or dies on consistency, and consistency is one of the few things that does not survive long without a policy behind it. If you are an SLT lead, a head of school, or a curriculum director weighing AI marking software for school moderation, the question you are really asking is whether a single platform can hold the standard across English, Maths, Sciences, Humanities and Languages without bending in any of them. This post is the short answer to that question, framed for the person who has to defend the decision to governors, parents and Ofsted.

GradeOrbit is built for that brief. The marking and detection workflow is the same for a single teacher in a classroom as it is for an entire department running a moderation cycle, and the institutional layer on top - per-teacher seats with unlimited use, school-wide onboarding, locked-down data handling - is what lets a policy actually take effect across every department at once.

What SLT Actually Need From Moderation Software

SLT do not need a marking tool that helps one teacher work faster. They need an instrument that produces the same answer in Year 9 English as it does in Year 11 History, and that produces it for the same reasons. That means three things in practice. First, every department uses the same engine - no shadow IT, no one head of department running their own thing on the side. Second, the engine applies the department's own rubric - not a generic one - so AQA, Edexcel, OCR, Eduqas and WJEC schemes are all honoured rather than averaged. Third, the trail of evidence is consistent: when a parent or governor asks how a grade was reached, the answer reads the same regardless of which subject it came from.

This is also where AI detection sits, because school moderation increasingly has an integrity layer on top of the rubric layer. If detection and marking come from two different vendors with two different data postures, the policy gets harder to write and much harder to defend. Having both inside a single platform - with a single data handling story - collapses that complexity.

Consistent Policy Across Every Department

The hardest part of a school-wide moderation policy is not writing it. It is keeping it consistent between, say, an English department that marks extended writing every week and a Maths department that mostly marks numerical work with occasional six-mark explain questions. The temptation is to write two policies - one for "essay subjects" and one for "exam subjects" - and watch the gap between them widen each year.

GradeOrbit handles both shapes in the same workflow. A department uploads its specific mark scheme once, and from that point every teacher in that department marks against the same rubric. A Head of Department running a moderation round sees the same kind of output regardless of subject: a suggested grade, a categorised feedback set tied to the rubric, and (where it matters) an AI detection likelihood score on student-typed work. That uniformity is what lets the policy stay one document rather than five.

For SLT mapping this to existing buying frameworks, our SLT guide to AI marking software walks through the diligence side in more depth.

One Price Per Teacher and Seat Planning

The institutional billing model matters more than it sounds. If every teacher in a school holds their own personal credit balance, you end up with a mess: some teachers run out in week 3 of term and stop using the tool, others hoard credits they never spend, and SLT has no visibility on actual usage. GradeOrbit's school plan replaces that with one flat price per teacher, billed annually, and every seat includes unlimited marking and AI detection. There are no credits to buy or ration, so a moderation round costs nothing extra. One subscription, one renewal, one place to see what got marked.

Seat planning sits next to that. SLT decides how many teachers should have access - usually a department-by-department rollout in the first term, then a school-wide expansion once the policy is bedded in. There is no per-teacher reactivation friction; adding a colleague is a few clicks. For SLT planning a September deployment around this, the guide above covers the practical sequencing.

Signatory Onboarding, School Email, URN Optional

School onboarding is where most institutional software loses an afternoon and four follow-up emails. GradeOrbit is designed to keep that interaction short. A designated signatory at the school - usually a headteacher or a deputy with budget authority - signs up using a school email address. The school's URN can be added during onboarding if you want it captured for your asset register, but it is optional rather than blocking. The signatory accepts the Data Processing Agreement, the school starts on a 30-day free trial with no card needed, and the rest of the staff team can be invited in by email straight away. The seat count is chosen when the trial ends, and nothing is charged until the school confirms.

The DPA is the bit that matters to your DPO. It is countersigned, version-controlled and held against the school's record, so when the next data audit lands you have one document to produce rather than a chase across email. For SLT building a tighter policy around academic integrity at the same time, pair this with how to write a school AI academic integrity policy.

Data Handling and Student Work

The single most important sentence in GradeOrbit's data posture is this: school accounts hold student work under a signed Data Processing Agreement, for a retention window the school sets. Scans are removed when a marking run archives (90 days by default, adjustable from 7 to 365 days), and pupil records are wiped at the end of the school's retention window (one to three years) or at offboarding. Only individual teacher accounts, outside a school agreement, discard work after marking. That is a deliberate design choice, and it is the one that lets SLT write a clean moderation policy with a short, factual line about where flagged scripts are kept and when they are deleted.

Because school accounts keep work against named pupils, the moderation policy can say exactly who sees flagged scripts and for how long, rather than relying on anonymisation. Combined with a stated retention window, that gives DPOs the answer they actually need when a parent asks where their child's work has been processed. For the longer version of this argument, see what happens to student work after AI marks it.

Talk to Us About GradeOrbit for Your School

If you are weighing AI marking software against a real school moderation policy - not in the abstract, but in terms of how it lands across English, Maths, Sciences and Humanities next term - the best next step is a conversation before you start the free trial. Different schools have different priorities: some lead with the marking workflow, some lead with detection and academic integrity, some lead with the consistency story for governors. We can talk through which entry point fits yours and what a sensible first-term deployment looks like.

Visit gradeorbit.co.uk to find out more, and get in touch from the homepage to discuss deployment for your school. We will walk you through the institutional plan, the DPA, and how per-teacher pricing and signatory onboarding work in practice.