Taptic's numbers give Gosai standing to make this argument. The platform recorded 3.6 million sessions in 2025, up from 3.1 million the year before, with SA Papers, its flagship product, drawing 5,000 students in a single 30-minute window at peak. 

Four new country editions launched last year alone, adding Eswatini, Namibia, Mauritius, and Seychelles to existing coverage in South Africa, Rwanda and Zambia. 

We covered Taptic's continental expansion in depth last November. This is not a startup speculating about the market. Gosai has watched how African students actually use education infrastructure at scale, and what he's observed points away from AI as a primary solution.

The edtech pitch is always the same: another platform, R50 to R200 a month, with AI that will change education in South Africa. Gosai gets approached regularly by organisations wanting to collaborate on exactly these kinds of products. His answer is consistently no, and the reason goes deeper than price.

The first problem is foundational. A recent study found that around 81% of Grade 4 learners in South Africa cannot read for meaning in any language. "AI can't teach a person to read," Gosai says. Before personalised tutoring bots or adaptive learning engines become relevant, there is a more basic crisis to address, one that no chatbot is equipped to solve.

The second problem is economic. South Africa has 64 million people. Of the roughly 21 million South Africans under 18, approximately 13 million depend on a Child Support Grant to survive. That grant pays R580 a month. An AI education platform at R100 a month consumes nearly a fifth of it.

A ChatGPT Pro subscription at R400 takes up more than two-thirds. The pricing that looks modest from the outside looks very different when it sits against the only income a household has. "It's almost impossible to budget for AI," Gosai says.

There is a third concern Gosai raises, and it's more uncomfortable than affordability. He is still a university student himself, and he says he has watched AI change how students approach thinking, including his own. The instinct to open ChatGPT before attempting a problem, to outsource the first draft of the essay, to skip the struggle of working something out. He has seen it widely enough among university students to treat it as a pattern.

In a country where millions of learners are already navigating significant disadvantages, compounding that with a tool that reduces the friction of thinking is not, in his view, a straightforward gift.

Gosai's answer is structural, and not something a product launch solves. South Africa needs a different curriculum, one designed to produce capable workers after Grade 9 or Matric rather than routing every student through a system built around university entry. Inventory management, healthcare assistance, sales and administration. The skills that a matric certificate currently does not reliably confer.

This is not an argument against the university, he says. It is an argument for a system that serves students whose circumstances, like walking hours to school and living off social grants, are not the circumstances around which the current curriculum was designed. We've explored the state of STEM education in SA separately, and the structural gaps Gosai describes are visible in that data too.

Gosai doesn't claim Taptic is the cure, and that honesty is part of what makes the argument credible. Three million sessions on past papers is valuable exam preparation. It is not a fix for reading comprehension, for infrastructure gaps, or for a curriculum that Gosai himself thinks needs fundamental redesign. Taptic is useful precisely because it works within the system that exists. The argument Gosai is making is that the system needs changing in ways that neither past papers nor AI tutors can substitute for.

You might also like: Read our feature on how Taptic's SA Papers platform scaled across Africa. See what the data says about AI adoption in South Africa in 2026 and where African businesses are actually deploying AI.