Colleges can now identify students at significant risk of premature withdrawal far earlier in the academic year allowing timely interventions. This is the first step in utilising Machine Learning (ML) for personalising student experience in this £6.4 billion pound industry.
The funding dilemma facing colleges
College education offers life changing opportunity. Funding is critical in maximising attainment, but this is not just linked to intake alone, instead requires students completing their course. Early leavers loose colleges 50% of their funding and therefore reducing this dropout is crucial. If students at risk could be identified early, then this would allow early intervention to help them stay on. A significant reduction in student attrition would not only be important for those individual students but could save colleges millions.
Our breakthrough research
Colleges collect large amounts of data on students’ behaviour, demographics, and attainment. This however produces large, complex and heterogeneous data sets. No individual risk factor on its own can identify the most at risk groups. Working with a large college serving 10,000 students annually, we used ML to explore different factors of premature withdrawal. In doing this, we created a model that identified a small group of students that were at a very high risk (>40%) as compared to most students in whom this risk was <5%. This model required the interplay of numerous demographic and behavioural data points, which was only possible with our innovative ML model.
How technology can identify, direct and automate
Having identified a clear at-risk group, we are now working with the college on ways to solve the dropout problem before it happens. Employing personalised digital assistants to help with appropriate course section, early identification and improved personalised learning based on tangible evidence can all lead to better outcomes. Decisions are no longer based on anecdotal insights or historical benchmarks. Instead, we can predict threats with increasing accuracy and take action.
Ground-breaking AI releasing up to 40% of staff time
ML is not the only technology that can impact the drop-out rates. When colleges deploy human parity digital assistants to give students 24/7 access to information, staff has less on their daily to-do lists. And no college has enough staff to do both: answer repetitive queries and provide personalised assistance to the most vulnerable students. When AI takes over most common questions, staff gains time to help those who are struggling.
Even though it’s an advanced technology, AI comes down to the most human needs. A need to be heard, find information quickly, solve a problem. With AI-driven technology, students who would normally get frustrated waiting for an email response can get the answer they need quickly and progress better at their training.
Our AI-driven student self-service releases up to 40% of staff time, enabling them to focus on high-priority areas.
AI-driven insights and ML model for resource allocation
Apart from answering questions, AI leads to long-term service improvement. By collecting data and identifying new queries, digital assistants help colleges become more in touch with the real needs of their students. How?
AI insights feed into the ML model, improving it! At present, ICS.AI can collect and standardise complex data ready for processing. Our current model has an accuracy of >70%, and we expect this to improve as we gain further insights from the personalised assistants. What is more, with more time and data, the model will discover more ways to help college students thrive.
The size of the opportunity
Student retention is an important part of colleges' operating models for budget projections. But there is little evidence they have been able to substantially improve retention and pro-actively intervene to improve course completion rates.
There are 234 colleges in England alone, preparing 1.7 million students at any one time (source). That is an average of 7300 students per institution. If we assume retention rate is fixed at 90%, and the average student brings in £5,000 then the annualised loss of earnings is £1.825 million. Moving retention rate up by a single percentage point is worth £182,500.
Providing the ultimate personalisation of the student experience, the AI technology and ML modelling can be just what the education sector needs, for both the colleges and the students.
ICS.AI are the UK's market leader in conversational AI solutions across the public sector. Based in Hampshire, we have customers across central and local governments, higher education and healthcare. As a Microsoft Gold Partner we have delivered channel shift and digital transformation with our proprietary SMART range of AI products.
Comments