I don’t usually devote a full post to a single webinar, but this one genuinely merited it. The session offered one of the clearest, most practice‑grounded explorations of how AI is reshaping assessment and feedback across FE and HE.
Teachermatic is an award-winning platform now with over 300 centres across the UK. Originally aimed at reducing the need for staff to learn prompt engineering. Developed by a team from across FE and HE. So staff did not need to learn how to set up all of this in Chat GPT, Co-Pilot, Gemini, etc and it still offers these distinct benefits.
I've been a fan since the project was piloted and then gained UFI seed funding.
In the Glasgow Region, some great work from Clyde College and Kelvin College around deploying and using Teachermatic. I hope someone from City of Glasgow College is reading this.
The system has now moved from an initial focus on creating learning materials and assessments to looking at how AI can support the actual assessment and feedback process. The new assessment tools aim to save staff time and improve the overall learning experience for students - a lot of good evidence that this is the case.
The platform is secure and works to make ethical and sustainable choices. AI in Assessment - is being used a great deal in formative assessment and feedback. Interesting - needs assent by students - workload reduction is a priority only for staff. Students are suspicious of AI being used for marking and feedback - they need reassured that AI improves the volume of feedback, consistency, and timelines and that the human element has not disappeared.
The assessment and feedback tool was originally set up to mark one piece of work at a time, but can now mark whole cohorts' submissions. The teacher still has a lot of agency. Teachermatic has had useful feedback from awarding bodies and regulators. AI can be used to support feedback. AI tools are there to speed up and enrich the feedback process. Ultimately, decisions and feedback come from the teacher. So support does not replace human marking.
Evidence that school teachers mark up to 8 hours, FE up to 10 hours, HE up to 8 hours per week on marking and feedback. Reminded that more timely feedback supports better learning. The Forgetting curve is real, and speed matters - the old two-week rule is too long. When that works, reminded that lots of academic teams miss even that two-week turnaround target.
The assessment feedback tool looks excellent and comes with national levels across the UK, including Scotland. Can cope with a rubric, even a simple unit description, and a range of student inputs, from graphical to handwritten and digital text. Comes with a whole range of different marking schemes - copes with up to 250 assessments at a time. Was particularly impressed by the ability of the system to cope with the grading of drawings and diagrams and handwritten submissions. Creates an annotated PDF with comments on work for tutor review - they can accept, delete, change, and add comments. Provides very rich feedback and consistency of grading. When complete, you download and share with the student.
Have developed a specialist platform for BTechs that gives feedback against the performance criteria in the unit standard.
The balance of workload now looks around 80% AI and 20% teacher
Saves around 144 hours of work per year
Cost is £30 per year, perhaps the actual input cost of one hour of marking.
Strong endorsement from staff on time saving, accuracy of grading, and richness of feedback, and positive feedback from students. Has gone a long way in improving work-life balance and improving learner satisfaction.
UK-based Azure servers do not train any LLM, retain all IP rights, and all submitted work is deleted, not retained. Jisc is a reseller and has done all due diligence. In the background on general AI, they work hard to remove risk and bias from the overall system, but on actual grading and feedback, the AI is calling on the rubric, etc, that the centre provides.
Reduces bias, improves the validity and reliability of grading
Works best up to HNC/D year one and two of degree provision, but is getting better at grading very extended writing that features in the final years of University learning. Introducing pre-marking - so students get feedback before final submission - and integrations with VLE platforms.
If you want to adopt, then lots of advance planning makes this a game-changer; staff and students need to be engaged from the start.
The SAFE Framework
One of the most valuable outputs of the wider project is the emerging SAFE framework, designed to help institutions that currently lack clear policy or guidance around AI in assessment.
SAFE emphasises:
Safeguarding: data protection and privacy
Augmenting: human judgement remains central
Fairness: reducing conscious and unconscious bias
Ethical practice: transparency and student understanding
It covers governance, consent, and procedures for both staff and students—practical scaffolding for institutions that want to move forward responsibly..
It is really worth reaching out to Peter and the team at Teachermatic and getting a free test account to try these tools out .
