Divisi, a South African agricultural technology startup that provides farm management software, is scaling an application designed to supply smallholder farmers with specific crop instructions. According to a report from Disrupt Africa, the system calculates agronomic variables for individual fields to replace regional climate estimates. After a user demarcates their property boundaries within the software, the platform evaluates the viability of more than 70 crops based on local soil and terrain data.
The underlying architecture relies on the Global Agro-Ecological Zones methodology developed by the Food and Agriculture Organization. The system processes temperature limits, water demand and soil chemistry to define planting windows and convert crop cycles into scheduled tasks. The operation depends on fixed parameters rather than generative artificial intelligence.
“It is deterministic – the same field and the same crop parameters return the same answer every time, and every score names the factor that limited it. No part of a recommendation is generated by a language model. That kind of analysis has existed for decades for anyone with a GIS workstation and a budget. What hasn’t existed is that computation on a smallholder’s phone, telling them what to do on a Tuesday,” Chikutukutu told Disrupt Africa.
The tool targets structural barriers within rural financial markets. Small agricultural operations often lack auditable production records and base decisions on broad weather forecasts. This lack of formal history prevents financiers from pricing risk, resulting in loan rejections. By logging crop development within the application, growers build a database to demonstrate their output to credit institutions.
“Then the records. Paper is fragile, admin is nobody’s favourite job, and most farmers finish a season without knowing which crop is profitable or whether they made money. That absence of information affects the entire value chain. An agri-financier looking at a farmer with no training credentials, no verifiable production history and no financial record has no way to price the risk, so they price it high, or they don’t lend at all,” Chikutukutu said.
The platform operates free of charge for farmers and prohibits the sale of user data. The company generates revenue through a corporate partnership program where food processors, input suppliers and insurance companies pay a fee for consented visibility into the fields of participating growers. Because the system runs on global public databases for climate and elevation, the software can function in new markets without requiring local data acquisition. The self-funded company reached 500 active users in its first month and is currently pursuing corporate customers and grant applications to support further expansion.
