This real-world case study illustrates how groundbreaking strides in AI-assisted feedback have the potential to speed up marking without compromising quality, increase education completion rates worldwide, and supply tailored feedback to learners at scale. Used as a formative assessment tool, this real-world example by a leading online learning organization saw their writing course completion rate rise by 7%, paper return rates drop by 20%, and 57% of their students describing the AI feedback as “Extremely” or “Somewhat Helpful”.


While education has moved on considerably from the traditional classroom-only model to the more online and advanced options, the challenges surrounding what makes learning impactful at scale remain a concern, with student engagement and completion rates still worryingly high across Europe and the rest of the world.


The key to truly effective learning lies in providing more tailored and one-to-one teaching methods, as evidenced by data like that shown in Bloom's 2 Sigma problem. However, the challenge of this high-quality teaching approach has always been how to do so at scale while also keeping costs down.


These challenges, sometimes referred to as the ‘iron triangle’, (quality, accessibility, and cost), have meant that many educational organizations can only typically solve two of the three challenges satisfactorily. The result of this is often over-standardized learning experiences and an overemphasis on knowledge recall rather than critical thinking. This understandably has a huge negative impact on learners and their educational outcomes, a situation that only becomes more pronounced for non-traditional learners, or learners from disadvantaged backgrounds.


From carers, students from nonaffluent families, or those with specific learning needs, the result of non-personalized, under-resourced, and low-quality learning environments is a community of students that fall behind their peers at a significant rate, often impacting overall completion rates and long-term employment opportunities.


Now, with the rise of widely accessible AI technology, the ability to create truly impactful learning experiences is, perhaps for the first time, possible. Arguably, its most powerful and needle-pushing application is as a tool for feedback via formative assessments.


In this real-world case study, we demonstrate how AI-assisted feedback has already resulted in a course completion increase of 7% and a paper return rate drop of 20% for a leading online education organization.


The AI-assisted feedback tool in question achieved a Quadratic Weighted Kappa (QWK) score of 0.88 - meaning it is on par with human scorers - and allowed the organization to create standards-aligned rubrics for auto-grading, enabling them to define their own grading criteria and save huge amounts of time grading without lowering quality. Further, the tool was also able to generate constructive and tailored learner feedback that could be instantly translated into a range of languages. This feedback was described as “Extremely” or “Somewhat Helpful” by 57% of the organization’s students and allowed them to prepare for exams, practice more critical thinking, and lower test anxiety.

Join us as we demonstrate how more personalized learning delivered successfully at scale might finally be within reach.