Ninety-one days is long enough for the excitement of launching to meet the discipline of daily use.
Ninety-one days is not a lifetime in agriculture, but it is long enough for the excitement of launching a platform to meet the discipline of daily use.
During its first 91 days in production, Umlimisi was not only made available to farmers. It was used alongside the daily operation of Ekukhuleni Farm.
That distinction matters. A feature can look complete during development and still fail when it meets a working farm.
These are the lessons that have become clearest.
Farmers do not wake up thinking about software modules. They think about what must happen today and what cannot be allowed to go wrong this week.
The most useful starting point is therefore not a dashboard full of data. It is a clear view of the work: what has been planted, what stage it has reached and which activities require attention.
Technology becomes useful when it reduces uncertainty around the next action.
A system can ask for a perfect record of every event, but a farm team has limited time and attention.
If capturing information feels like a separate administrative project, it will become inconsistent. Records need to fit the natural rhythm of planning, completing work, observing problems and recording harvests.
The lesson is simple: every piece of information requested must earn its place.
A crop photograph can help identify a possible problem, but a single image does not explain the entire production history.
The crop, field, planting date, recent activity, symptoms and interventions all create context. Diagnosis becomes more valuable when it is connected to this wider record.
AI can support the farmer, but it must not create false confidence. Uncertain or serious cases still require professional agricultural support.
Ekukhuleni Farm has an on-the-ground manager while one founder is based in the United Kingdom.
The platform should not become a mechanism for watching every action. It should carry routine information so the manager does not have to report the same facts repeatedly and the owner does not have to keep asking.
Good visibility creates trust and more focused conversations.
A beautiful harvest can still hide an unprofitable production cycle.
The field record must eventually connect activity to inputs, costs, harvests, sales and profit. Farmers need to know not only what grew, but whether it created a sustainable return.
That connection is essential if Umlimisi is to help turn farms into stronger agribusinesses.
Banks, insurers, cooperatives, extension services and government programmes can all benefit from better agricultural information.
But the platform cannot begin by asking, “What data can we extract for institutions?”
It must begin by helping the farmer. Only then, with clear permission, can selected information support another organisation.
The farmer should not carry the burden of data collection without receiving value in return.
Operating the platform on a real farm creates a discipline that desk-based planning cannot reproduce.
The farm reveals where language is confusing, where tasks are unrealistic, where information is missing and where a feature adds genuine value. It also prevents us from building capabilities simply because they sound impressive.
The farm keeps asking the most important question: does this help?
Ninety-one days does not prove that every problem has been solved. It proves that Umlimisi has entered the environment where the right problems can be discovered.
We will continue to build in the same way:
Discover what is true. Build from it. Measure it. Improve it. Repeat.
That is how Umlimisi will grow—from a platform running on one real farm into infrastructure that can support farmers and agricultural organisations more broadly.
Follow the Umlimisi story as we continue building.