Case Study
Queen Mary University of London transforms assessment for the AI era
How QMUL is redesigning assessment for the AI era improving student outcomes, increasing engagement, and scaling structured, learning-centred assessment across the institution.
Share with colleagues
Download the full Case Study
Take an in-depth look at how educators have used Cadmus to better deliver assessments.
Highlights
82.2%
Positive student experience with Cadmus
+4%
Increase in average student grades
-40%
Reduction in first-time failure rates
QMUL is rethinking how assessment supports student success in an increasingly complex, AI-enabled landscape.
Following the rapid shift to online assessment, the university faced growing pressure on multiple fronts: rising academic misconduct, declining NSS scores for assessment and feedback, and new uncertainty introduced by generative AI. Traditional assessment models, often fragmented and submission-focused, were no longer fit for purpose. QMUL recognised that the answer wasn’t stricter controls, but better design.
In partnership with Cadmus, the university began redesigning assessment as a structured, learning-centred process, one that supports students throughout the journey, not just at the point of submission.
From submission to structured learning
Rather than relying on one-off, high-stakes tasks, assessments were rebuilt as guided experiences. Students were supported step-by-step from understanding the task and engaging with literature, through to drafting, feedback, and final submission. This shift fundamentally changed how students approached their work: not as a last-minute task, but as an active, iterative learning process.
“Providing detailed information of what is required in the assessment directly alongside the assignment itself is very helpful.”
At the same time, assessments were redesigned to be more authentic, grounded in real-world application, and focused on developing critical thinking and problem-solving skills.
For educators, this brought a new level of visibility. Instead of only seeing the final submission, they could understand how work was developed over time, enabling earlier intervention, more meaningful feedback, and greater confidence in the authenticity of student work.
“Cadmus encourages detailed, pedagogical instructions, improving student understanding and reducing assignment-related queries.”
Scaling from pilot to institution-wide impact
What began as a targeted pilot quickly gained traction. In its first year, Cadmus was introduced across 20 modules, reaching 1,435 students. By year two, this had expanded to 63 modules and nearly 5,000 students. Today, the initiative spans around 96 modules, with over 5,000 students engaging in Cadmus-enabled assessments.
This growth reflects more than adoption, it signals a broader institutional shift toward structured, digital assessment practices.
Measurable improvements in student success
As assessment design evolved, so too did student outcomes.
- QMUL recorded a 4% increase in average grades alongside a 40% reduction in first-time failure rates, pointing to meaningful improvements in academic performance.
- Student experience also improved, with 82.2% reporting a positive experience and 75.9% actively engaging with feedback, a critical indicator of deeper learning.
Together, these results highlight a clear shift: when assessment is designed to guide learning, engagement increases and outcomes follow.
A better experience for educators and students
For educators, the impact was immediate. Assessments became clearer, more structured, and easier to manage, with fewer student queries and stronger alignment between teaching and evaluation.
“It’s a lot easier to invigilate with Cadmus. Student feedback is very positive: they like the timer, word count, it puts them in the right mindset.”
For students, the experience shifted from uncertainty to clarity. With expectations embedded directly into the assessment and opportunities to draft and improve their work, students reported greater confidence, stronger engagement, and a deeper focus on understanding — not just completion.
“Working in Cadmus taught me to use literature search, the use of guidance for planning, and be succinct when making points.”
“I think it should be used for all module assignments, everything is in one place, it’s easier for students.”
Designing assessment for what comes next
QMUL’s work reflects a broader change taking place across higher education. In an AI-enabled world, assessment can no longer rely on traditional models of verification alone. Instead, it must be designed to actively support learning, surface evidence of thinking, and provide educators with confidence in student work.
By embedding structure, scaffolding, and visibility into the assessment process, QMUL is building a model that is not only more effective today, but resilient for the future.
Category
Student Success
