South African startup Exten AI has developed an artificial intelligence platform that enables entrepreneurs with no coding background to build functional software applications through simple descriptions written in everyday language. The tool automatically generates a complete system, including an isolated database, file storage, secure user authentication mechanisms, and the environments required to host and run autonomous AI agents.
According to Exten AI co-founder Luncedo Simelane, the platform converts instructions written in natural language into ready-to-use applications. Unlike many text-to-code tools, it does not stop at creating a visual interface — it also provisions the underlying infrastructure each application needs to operate.
“It is essentially a digital technical co-founder for an entrepreneur who has no technical background but a strong software idea, capable of delivering a production-ready structure in under 30 minutes,” Simelane said.
Before committing fully to software architecture, Simelane taught himself to code and left his law studies to focus on technology development.
It was during that journey, he says, that he recognised technology infrastructure limitations as one of the principal barriers to innovation across Africa. Simelane describes the challenge as a “technology infrastructure problem” — the gap between conceiving an idea and bringing it to technical life.
“Global text-to-code platforms mainly serve existing developers, allowing them to code faster, or they are limited to generating superficial visual interfaces,” he explained.
“Global text-to-code platforms mainly serve existing developers, allowing them to code faster, or they are limited to generating superficial visual interfaces” — Luncedo Simelane, co-founder, Exten AI
For Simelane, solo entrepreneurs and small businesses face particular difficulties when they need to implement backend systems, structure relational databases, or ensure authentication and security mechanisms are in place. “Exten AI solves exactly that problem, functioning as an automated technical partner,” he added.
In competitive terms, the co-founder positions the platform between global tools that allow users to build interfaces through natural language instructions but do not provide full databases, and enterprise low-code development platforms such as Microsoft Power Apps, which he regards as too expensive and rigid for builders in emerging markets.
The startup is currently self-funded by its founders. An initial proof-of-concept received backing from Professor Mamokgethi Phakeng, former deputy vice-chancellor of the University of Cape Town, during the Sunday Sessions programme.
More recently, Exten AI launched a pilot project in partnership with the Centre for Entrepreneurship and Rapid Incubator at the University of Fort Hare. Participating student entrepreneurs will use the platform to develop early versions of their products and businesses.
Simelane said demand for the platform quickly outpaced initial server capacity. The first public launch recorded more than 100 sign-ups, prompting the company to move to a closed beta phase with a restricted group of users in order to ensure infrastructure stability.
Early users have already developed functional applications, among them LedgerAI, designed to automate accounting processes; EduAgent, built to assist teachers with lesson planning; and LuminaAI, a personalised communications dashboard.
Exten AI operates on a software-as-a-service (SaaS) model with tiered usage levels. “While building and testing applications is straightforward and frictionless, users pay a predictable monthly subscription when their applications are made publicly available,” Simelane explained.
Pricing varies according to server usage, data storage, and the use of AI agents. The company acknowledges that early revenues have been modest, given that its focus during the closed testing phase has been on optimising its architecture and improving unit economics — that is, the relationship between revenues and the costs associated with each customer or application.
Scalability has been one of the startup’s foremost challenges. According to Simelane, simultaneous use of the platform by dozens of people — each deploying AI to create, configure, and implement complex cloud database structures — initially caused infrastructure bottlenecks.
Source: Diário Económico
Original article: https://www.diarioeconomico.co.mz/?p=549028
