Comparison

Human evaluators or AI marking?

Automated marking is fast and cheap, and for some question types it is genuinely fine. For extended responses and method-based working, a qualified human still decides the mark. Here is where each one holds up.

Where automated marking genuinely wins

  • Instant turnaround. Seconds, not 48 hours.
  • Cost at volume, a few pence a script rather than a few pounds.
  • Objective, closed questions: multiple choice, single numerical answers, recall.
  • Unlimited retries, which suits low-stakes drilling between real papers.

We'd rather you knew this before you bought anything. The table below is scored the same way.

Head to head

Human evaluators vs Automated marking

Ten criteria, scored in both directions.

Turnaround
Time from upload to feedback
Automated marking:Seconds
Human evaluators:48 hours
Advantage: the alternative
Cost per script
Marginal cost
Automated marking:Pennies
Human evaluators:From £7.99
Advantage: the alternative
Closed, objective questions
Recall and single-answer items
Automated marking:Reliable
Human evaluators:Reliable
Roughly equal
Method marks in working
Credit for a correct method with a wrong final answer
Automated marking:Unreliable, non-linear handwritten working is frequently misread
Human evaluators:Marked line by line as an examiner would
Advantage: GCSE Feedback
Extended writing
English essays, 6-mark science responses
Automated marking:Produces a band, rarely a defensible one
Human evaluators:AO-by-AO commentary and band justification
Advantage: GCSE Feedback
Handwriting and diagrams
Real scripts, crossings-out, graphs, annotated figures
Automated marking:Error-prone
Human evaluators:Read as a human examiner reads it
Advantage: GCSE Feedback
Mark scheme judgement
Applying 'or equivalent' and allowing valid alternatives
Automated marking:Literal, valid alternative answers get marked wrong
Human evaluators:Judgement applied, as in the real series
Advantage: GCSE Feedback
Actionable next steps
What to do differently next paper
Automated marking:Generic
Human evaluators:Three specific prioritised actions per paper
Advantage: GCSE Feedback
Consistency across a cohort
Comparability between classes
Automated marking:Perfectly consistent, including consistently wrong
Human evaluators:Moderated sampling on every batch
Advantage: GCSE Feedback
Trust with parents and schools
Whether the grade can be defended
Automated marking:Rarely accepted for reporting
Human evaluators:Marked by evaluators who mark for the boards
Advantage: GCSE Feedback

A tick marks the option we think wins that row. Three of them don't point our way.

Automated marking suits

  • Multiple-choice and recall drilling
  • High-volume low-stakes practice
  • Instant self-checking between real papers
  • Zero budget per script

Human marking is needed for

  • Extended writing and essay bands
  • Method marks in handwritten working
  • Mock grades used for reporting or intervention
  • Anything a parent or department will act on
In short

Where we stand

Every mark and every comment on a GCSE Feedback script is written by a qualified human evaluator. That is the product, and it does not change.

Software has a place in our operation, routing uploads to the right specialist, checking a photo is legible, flagging a script for moderation. It never decides a mark, and it never writes the feedback a student reads.

If you need a score in ten seconds for a set of multiple-choice questions, an automated tool is the right tool. If the mark has to hold up in front of a parent, a head of department or the student themselves, a person needs to have read the paper.

FAQ

Questions people ask before deciding

Do you use AI to mark papers?
No. Every mark and every comment is written by a qualified human evaluator. Software helps us route uploads, check legibility and flag scripts for moderation, it does not award marks or write feedback.
Isn't automated marking accurate now?
On closed, objective questions it is fine. On handwritten working and extended responses it still misreads non-linear layouts, penalises valid alternative answers, and produces band judgements it cannot justify against the mark scheme.
Why does it take 48 hours then?
Because a person reads the script, applies the board's mark scheme question by question, annotates it and writes the report. That is the time it takes to do properly, and a proportion of every batch is moderated on top.
Who are your evaluators?
UK-based qualified subject teachers, many of whom also mark for AQA, Pearson Edexcel or OCR during the summer series. Every script goes to a specialist in that subject.

Judge it on one paper

One full paper marked free by a qualified evaluator, returned in 48 hours. No card details.