South Africa’s private tutoring market is often discussed as a sign of parental ambition. That is only partly true. In many households, tutoring is not about chasing an edge. It is a response to a basic gap in the education system: learners need more individual feedback than crowded classrooms can realistically provide.
The country does not lack educational content. Learners have textbooks, worksheets, exam papers, online videos, school portals and WhatsApp groups full of study material. What many do not have is someone who can look at a wrong answer and identify the missing concept behind it.
That distinction matters. More content does not automatically produce better learning. A learner can watch a lesson, copy the method, memorise a procedure and still fail to understand why it works. The real value of good tutoring is not simply extra time. It is diagnosis, explanation, correction and repetition.
This is why tutoring has become so attractive to parents. A tutor can slow down. A tutor can ask the learner to explain their thinking. A tutor can notice that a Grade 9 algebra problem is really a Grade 6 fractions problem that was never resolved. A tutor can give feedback while the learner is still mentally close to the mistake.
That kind of attention is difficult to provide at scale. South African class-size research cited in recent education analysis shows large classes and rising learner-educator ratios, with some early-grade classrooms carrying more than 40 learners. In that environment, even capable teachers face structural limits. A teacher may know exactly what good feedback looks like, but still not have enough time to give it to every learner.
The result is a quiet shift in responsibility. Families who can afford it buy the feedback loop privately. Families who cannot are left to rely on the classroom alone, even when the learner needs more.
This creates an uncomfortable equity problem. Tutoring can help close learning gaps for individual children, but the market can widen inequality when access depends on household income. The supplied research for the Zanetti AI Institute’s recent white paper found private tutoring costs ranging from modest hourly rates to intensive matric support costing thousands of rand a month. Against a median annual household income of about R95,770 in 2023, even a seemingly moderate tutoring bill can be out of reach for many families.
The issue is not that tutoring is bad. The issue is that feedback-rich learning support has become priced like a private advantage.
South Africa’s national learning data reinforces the seriousness of the problem. PIRLS 2021 reporting indicated that about 81% of Grade 4 learners could not read for meaning. TIMSS 2023 reporting again showed severe mathematics and science underperformance. At the same time, South Africa has reported historically high matric pass rates. These facts should not be used to dismiss matric achievement. They should, however, remind us that progression through the system does not always mean durable understanding.
The education sector therefore needs to ask a sharper question. How do we make individual academic support more widely available without lowering quality, bypassing teachers or replacing proper schooling with shortcuts?
This is where AI academic support deserves careful attention. Not because it is a magic answer, and not because every chatbot should be allowed near a child’s learning journey. It deserves attention because the core problem is a scale problem. Many learners need explanation, practice and feedback more often than the system can currently provide.
A well-designed AI tutoring system can, in principle, repeat explanations patiently, adapt practice to the learner’s level, support learning after school hours and reduce the cost of access. But the phrase “well-designed” is doing a lot of work. Generic AI is not enough. For academic support, the system must be curriculum-grounded, age-appropriate, language-aware and subject to human oversight.
For South Africa, curriculum alignment matters. A learner does not need an impressive answer from the internet. They need an explanation that fits their grade, subject, syllabus and assessment expectations. Language matters too. Many learners are trying to master concepts in a language that is not the one they use most naturally to think. Translation alone is not the same as understanding.
The responsible path is neither hype nor rejection. Schools, funders and education partners should pilot AI academic support where the learning purpose is clear, the curriculum boundaries are defined and the outcomes can be measured. Usage is not enough. Time on platform is not enough. The real test is whether learners understand more, practise better and receive feedback that helps them correct the right thing.
Private tutoring has shown us what many learners need. The next challenge is to make that kind of support more accessible, more accountable and less dependent on a household’s ability to pay.
About LearnerFlow
LearnerFlow is relevant because its stated design addresses real South African needs: CAPS and IEB alignment, Grade 4 to Grade 12 support, multilingual access, voice-based explanation, past-paper practice and memorandum-aware marking.
