Marking aid is an AI-powered assessment feature within Aptem Enhance, the AI feature set inside Aptem Apprentice. It is designed for training providers who want to reduce the time tutors spend marking written work, improve the consistency of feedback, and give learners faster results. This guide explains how marking aid works in practice, from creating exercises to submitting, marking, and managing multiple attempts.
This article covers: how marking aid fits into the Aptem workflow, the learner experience, the tutor marking workflow, how the classification AI improves its suggestions over time, and how to manage exercises across programmes.
What marking aid does and why it matters
Marking written work is time-consuming, and the speed of feedback directly affects learner progress. When feedback is slow, learners wait, planned end dates slip, and some learners disengage. Marking aid addresses this by supporting tutors with AI-generated feedback suggestions and structured marking tools, so they can focus on the parts of their job that require professional judgement rather than repeating the same standard comments across similar submissions.
The key principle is that the tutor remains in control. Marking aid does not mark work automatically or replace the tutor’s decision. It suggests, the tutor confirms or overrides, and the classification AI refines its suggestions over time as more submissions are marked against the criteria your team has defined. Providers who have embedded marking aid report saving up to 30 minutes per submission, alongside more consistent feedback and cleaner version control across their curriculum.
Marking aid can be used for any written work: apprenticeship standards, qualifications, diplomas, certificates, functional skills mock papers, wider learning activities such as British values and Prevent, and first-day-of-learning induction tasks. It is not locked to a specific standard or framework.
How marking aid sits within Aptem
Marking aid is part of Aptem Enhance, which sits within Aptem Apprentice. It works through two integrated components: a marking interface (the embedded marking platform) and the Aptem marking aid management screen, which gives administrators and quality teams a centralised view of all exercises across programmes. Marking aid uses classification AI, which works by sorting submissions against predefined criteria rather than generating new content. This makes its suggestions transparent and explainable, supporting Ofqual’s expectations around transparency in AI-assisted marking.
Each exercise is linked to a learning plan component, which is what the learner sees and interacts with in their programme. The component type can be a task assignment, digital learning, or other standard component types, so the marking aid exercise fits naturally into the existing structure of the learning plan. Marking aid exercises are managed from a dedicated marking aid administration screen, giving curriculum and quality teams more direct control over exercises without navigating through programme configuration.
Creating a marking aid exercise
Setting up a marking aid exercise typically takes 15 minutes or less. When creating the exercise itself, you build each question in a rich text editor that supports text formatting, images, PDFs, tables, and hyperlinks to external resources. Questions can include reading passages, indicative content, links to further learning, or detailed task briefs, giving learners everything they need within the submission interface. If you already have existing workbooks or assignment briefs, content can be dragged and dropped or copied directly into the exercise builder.
You can create exercises from scratch, by copying from existing materials, or by uploading mock exam papers (particularly useful for functional skills reading and writing papers). Word counts can be applied to individual questions. Once questions are created, they can be edited but not added to retrospectively, so the recommended approach is to finalise all questions before publishing the exercise.
After creating questions, you have three options: publish the exercise as-is without rubrics or feedback, add rubrics and feedback banks before going live without pre-training the AI, or add rubrics and feedback and train the AI using test learner submissions before launch. Larger providers typically use the third approach, running test submissions to build up AI confidence before live learners are enrolled.
Building rubrics and feedback banks
The marking accuracy and efficiency of marking aid depend on two components: rubrics and feedback banks. These can be built before or after exercises go live.
A feedback bank is a library of reusable feedback items attached to a question. When marking, tutors can select items from the bank to apply to a submission. The classification AI identifies which items are most appropriate for different types of response, and over time begins surfacing the most relevant feedback automatically. Feedback items are categorised as positive or constructive, and can be organised into groups, for example separating ‘what went well’ from ‘how about we try’. The feedback bank belongs to the exercise, so any tutor marking that exercise can draw from the same pool, supporting standardisation across teams.
A rubric is a structured mark scheme applied to a question. Rubrics assign marks to specific criteria, such as response to the command verb, use of relevant knowledge, or contextual awareness. Rubrics can use discrete grading (marking in full grades, such as pass, merit, distinction) or continuous grading (allowing partial marks within a band). Multiple rubrics can be applied to a single question. When AI grading is enabled on a rubric, the classification AI begins suggesting grades as more submissions are marked against those criteria.
For multiple-choice style questions and functional skills mock papers, the feedback bank approach works particularly well: tutors add correct and incorrect answer feedback, and the quiz becomes close to self-marking once the classification AI has processed a sufficient number of submissions.
Rubrics and feedback banks are built by tutors and curriculum staff who know the subject. The classification AI does not generate feedback content; it surfaces the most relevant items from the bank your team has built, based on how submissions compare to the predefined criteria.
The learner experience
Learners access marking aid exercises through their learning plan in the same way they access any other component. The component appears with instructions, a complete-by date if set, and any off-the-job hours associated with it. The learner selects ‘submit exercise’ to enter the marking interface.
Inside the exercise, learners see the assignment description (which can include detailed task briefs, referencing requirements, word count targets, and links to supporting resources), followed by each question in turn. The response editor supports bold and italic formatting, bullet points, tables, hyperlinks, images, and PDF uploads. Learners can answer one question at a time, save their progress, and return later without losing their work.
When a learner submits, they record their off-the-job training hours as they would for any other evidence submission in Aptem. The submission then appears in the tutor’s marking queue.
Marking aid works across devices. Learners who primarily use mobile phones can submit and view feedback on the same device, removing the need to download, complete, and re-upload workbooks.
How tutors mark submissions
Once a learner has submitted, the tutor can access the work from their ‘requires marking’ screen in Aptem or by navigating to the learner’s profile. Selecting ‘mark exercise’ opens the marking interface.
For each question, the tutor sees the learner’s answer alongside any rubrics for that question. If the classification AI has processed enough previous submissions to begin making confident suggestions, it will suggest a rubric grade and relevant feedback items automatically, along with a confidence percentage. The tutor reads the answer, checks the AI’s suggestions, and accepts or overrides them.
Alongside rubric grades, tutors can select items from the feedback bank to attach to specific parts of the answer. Feedback can be applied to highlighted text, specific paragraphs, or images and PDFs attached as supporting evidence. One-time comments are also available for feedback that is specific to that individual learner, such as a spelling correction or a contextualised note about their employer. One-time comments do not feed into the feedback bank or train the AI; they remain private to that submission.
When marking is complete, the tutor either accepts the submission or refers it back to the learner with feedback. Referred work generates a notification to the learner, who can then view all the feedback given, see their rubric scores, and begin a new attempt.
Multiple attempts and the audit trail
When a submission is referred, the learner’s original attempt is preserved in full. The learner begins a new attempt from within the same component and can see all previous feedback when doing so. Some providers ask learners to add new content in a different colour so the tutor can quickly identify what has changed between attempt one and attempt two.
The evidence summary in Aptem shows all attempts for each component, labelled as attempt one, attempt two, and so on. Internal verifiers and quality teams can view each attempt separately, see how the learner’s work and scores progressed, and use this as evidence of learning and development during Ofsted inspections or external quality assurance visits.
Marked submissions feed into Aptem’s portfolio export. It will also be possible to export a version of the portfolio without the tutor feedback included, to meet end-point assessment organisation requirements where annotated feedback should not be submitted alongside the learner’s work.
Managing exercises across programmes
The marking aid management screen gives administrators a full view of all marking aid exercises across delivery programmes and subprogrammes. From here, exercises can be searched, filtered by group, employer, or primary tutor, edited, and cloned.
Cloning an exercise copies the questions, feedback banks, rubrics, and any classification AI confidence built to date into a new version, which can then be linked to a different programme or component. This is useful when the same or similar assignment is used across multiple standards, for example a project management or interpersonal skills exercise that appears in team leader, operations manager, and business administrator programmes. The clone can then be edited to adjust question wording or command verbs for the specific level or standard without affecting the original.
New exercise versions can also be created at the individual learner level, for example to add scaffolding questions or simplified wording for learners who need additional support, or to add employer-specific questions requested by a particular organisation. Other learners on the same programme remain on the original version.
When programme changes are made, the sync function in Aptem allows tutors and administrators to push updated exercises to specific learners or cohorts already on programme, without requiring each learner to be re-enrolled.
Frequently asked questions
Does marking aid replace the tutor’s judgement?
No. Marking aid supports tutors by surfacing relevant feedback and suggesting rubric grades, but every marking decision is confirmed by a tutor. The classification AI refines its suggestions as more submissions are marked, meaning its accuracy improves over time and reflects the marking standards your organisation has defined. Tutors remain responsible for all marking outcomes.
What types of assessment can marking aid be used for?
Marking aid can be used for any written assessment: apprenticeship standard work, qualifications and diplomas, functional skills mock papers, wider learning activities, and induction tasks. It is not limited to a specific standard or framework. Many providers use it across their full curriculum, including British values, Prevent, and health and safety induction tasks that are common to all learners.
How does the AI improve over time?
Marking aid uses classification AI, which learns which feedback items and rubric grades are most appropriate for different types of response as more submissions are marked. The more your team marks, the more accurately it surfaces relevant suggestions – always within the criteria your team has defined. Providers can accelerate this by running test learner submissions before going live, which builds up the AI’s confidence before real learners are enrolled.
Can marking aid be used for functional skills?
Yes. Marking aid is well-suited to functional skills mock papers for reading, writing, and mathematics. Reading and writing papers can be built by copying text from awarding body materials into the exercise builder. Once the feedback bank and rubrics are in place, marking becomes significantly faster. Providers should still ensure learners practise using awarding body platforms directly, as that prepares them for the tools used in the actual exam.
What happens if a learner needs a different version of an exercise?
It is possible to create a new version of an exercise for an individual learner or a specific cohort, for example to provide extra scaffolding for SEND learners or to include employer-specific questions. Other learners on the programme remain on the original version. The sync function in Aptem can push the new version to selected learners without affecting anyone else.
How does marking aid support quality assurance?
The marking aid management screen gives quality teams a centralised view of all exercises, learner progress, and marking activity. Internal verifiers can access each attempt for each learner and review marked work. The full attempt history, including rubric scores and feedback for each submission, is available in the evidence summary and feeds into Aptem’s portfolio export, supporting both internal quality assurance processes and external quality assurance visits.
To see marking aid in action, book a demo or watch our detailed training series below.
