ORT South Africa Launches Trilingual AI Study Companion

ORT South Africa has introduced a trilingual artificial-intelligence study companion for participants in its big-data analytics programme, giving learners access to curriculum-based support in English, isiZulu and isiXhosa.

The tool is called Funda Nam, meaning “study with me” in isiXhosa. It has been trained on the programme's own learning material and is intended to complement physical training rather than replace instructors. Learners can ask questions, review previous anonymous questions and receive work-readiness support outside scheduled sessions.

ITWeb reported that ORT South Africa developed the service with GAS Marketing Automation. The underlying big-data programme began in 2024 with support from the Victor Daitz Foundation and has reached more than 200 young people in KwaZulu-Natal.

The language design is the most important part of the launch. Many educational AI tools are technically available in South Africa but work most reliably in English. A learner may understand a difficult concept more clearly when it is explained in the language used at home, especially when the subject itself contains unfamiliar technical terminology.

Supporting three languages does not automatically guarantee equal quality. ORT will need to test whether answers remain accurate across languages, whether technical ideas are translated consistently and whether the system understands the ways learners actually move between English, isiZulu and isiXhosa. Usage data should be reviewed by educators and language specialists rather than treated as a purely technical performance measure.

Grounding the assistant in a defined curriculum is a sensible safeguard. A general chatbot may produce a fluent answer that conflicts with course material or introduces concepts beyond the learner's level. A narrower system can point students back to the relevant lesson and make it easier for teachers to evaluate the quality of responses.

Human escalation remains essential. Learners should be able to flag unclear or incorrect answers and know when a question has been referred to an instructor. The system should also distinguish between helping someone understand an exercise and completing assessed work on their behalf.

The programme fits into a wider shift in African education. Nigeria is developing a national AI roadmap for universities, while governments and companies are funding large-scale skills programmes. Funda Nam operates at a smaller scale, but its focus on language and a specific curriculum may make its results easier to observe.

Access conditions will influence those results. A 24-hour service is only useful if data costs, device requirements and login processes are manageable for the intended learners. ORT should publish basic information about availability, privacy, retention of learner questions and whether conversations are used to improve the model.

There is also an opportunity to measure more than engagement. Useful indicators would include course completion, understanding of difficult concepts, confidence across languages, employment outcomes and the frequency with which human support is required. Those measures would show whether the assistant is improving learning rather than simply increasing interaction.

Funda Nam is not a substitute for well-supported teachers or sustained investment in education. Its promise lies in extending structured help between sessions and meeting learners in languages that reflect their daily lives. If the organisation evaluates it openly, the project could offer practical evidence about what locally adapted educational AI should look like.

Source: ITWeb.