AI-powered SET tool flips the script on student support

South Africa’s education system faces a stark dropout crisis, with attrition often building unnoticed over years until it’s too late. Despite the 2025 matric class achieving a record 88% pass rate, this milestone masks a harsh truth: Only 778 000 full-time learners sat the exams out of 1.2 million who started Grade 1 in 2014.

The reality is that schools often react too late, relying on exam failures or visible crises. Typically, interventions kick in only after exam failures or obvious disengagement, driven by a lack of meaningful, ongoing data that forces reliance on retrospective test scores as the main signal.

On top of that, teachers face training gaps in handling behavioural and wellbeing issues; schools have limited access to specialist expertise; accountability and the evaluations for interventions are weak; and support processes are time-heavy amid already stretched workloads.

Although traditional education systems have made progress in thinking about student support, far less progress has been made in implementing it both consistently and at scale. These trends indicate that learner attrition typically occurs well before the matric year, often after extended periods of disengagement or underperformance. As a result, interventions that rely solely on end-of-year or exam-based indicators may occur too late to provide meaningful support.

Early identification of learner challenges and difficulties as well as structured followup support are therefore critical components in addressing learner dropout and improving retention outcomes.

Koa Academy, a forward-thinking online school, has developed the SET Student Support Tool: a predictive, AI-supported system that spots struggling students ahead of when traditional methods detect issues. It analyses real-time learning management system (LMS) data on attendance, performance and engagement alongside teacher observations across three domains: the Student, learning Environment, and Task (SET). This breakthrough empowers teachers with actionable insights, transforming reactive support into proactive, scalable intervention.

Koa co-founder and principal Mark Anderson reframes AI’s role: not as a human substitute but as a catalyst for better human delivery. “AI changes how we design and deliver support, ensuring the right people connect with learners sooner and more meaningfully,” he emphasises.

Feedback from the 2025 AI in Education Forum 2025 from educators lauded SET for closing the ‘recognition-to-response’ gap with practical, accessible guidance.

The hidden gap in student support

Experienced teachers in both online and traditional schools often spot learner difficulties early: falling grades, withdrawal or inconsistent participation. The more significant gap often emerges after that recognition. Even when a concern is identified, teachers may not have the specialist training or tools to determine why it’s happening and determine which interventions are most appropriate. As a result, the challenge is less about awareness and more about knowing how to respond effectively.

Koa’s SET tool bridges this divide. It pairs real-time diagnostic data with structured guidance, shifting teachers from ‘something’s wrong’ to ‘here’s what to do next’. Analysing engagement patterns across student traits, learning environments and task demands, it prevents escalation and fosters retention.

“Strong student support isn’t just beneficial – it’s transformative. It builds intellectual security so learners embrace risks and persist through challenges; it boosts engagement and leads to improved academic outcomes, as University of Amsterdam research confirms; safeguards mental health and curbs risk behaviours per the 2023 CDC study; and fosters self-regulation through metacognition, according to the Education Endowment Foundation,” notes Anderson.

How the SET Tool works

The SET Tool leverages data from Koa’s LMS, collated daily into accessible student dashboards for teachers and parents. It tracks attendance, performance and engagement metrics automatically. Early triggers are flagged as potential risks that prompt the initiation of a SET report.

Teachers capture nuanced observations on the student (learning experience and challenges), environment (physical and socio-emotional context), and the task (how learning is presented and what is required of the learner).

AI-supported analysis, using purpose-built algorithms, combines data to reveal patterns and recommend tailored interventions.

Lastly, historical insights refine future outputs, enhancing accuracy over time.

Unlike automated AI replacements, SET acts as a human-centred decision-support layer that is teacher-led, judgment-respecting and relationship-focused.

Beyond detection: Actionable, predictive power

Traditionally, schools rely on reactive signals (exam failures, disciplinary issues or crises), delaying support until damage is done. SET’s predictive analytics intervene proactively, surfacing subtle trends ahead of traditional methods.

It integrates social-emotional and cognitive factors holistically. This enables earlier, targeted responses to issues like emotional distress, self-management gaps or mismatched tasks, directly tackling South Africa’s high secondary dropout rates.

Koa Academy’s AI-supported approach is designed to enhance student support processes and scale teacher expertise, enabling consistent, high-quality support without removing the human relationships at the heart of learning. It functions as a decision-support layer that helps educators interpret patterns, identify potential areas of concern earlier and respond with greater clarity and confidence. All interventions remain teacher-led, with professional judgement and contextual understanding playing a central role.

By strengthening how early signals of disengagement, underperformance or overwhelm are identified and responded to, this approach supports earlier and more structured intervention during the years when learners are most at risk of exiting the schooling system.

In a context where attrition often occurs well before matric, this focus on timely and effective support helps address some of the underlying factors that contribute to dropout, rather than reacting only once these challenges have already escalated.

As Koa Academy demonstrates, schools can break this cycle with AI-supported systems that empower teachers for proactive interventions.