NextBillion - A WDI Publication

There is a dominant mental model of AI for agriculture. According to Sheena Raikundalia at Kuza, the pitch goes something like this: “What if every farmer had a personal AI advisor telling them what to grow, when to plant and how much fertilizer or other inputs to apply?” That model is sleek, frictionless, and built on the assumption that farmers are operating alone with the algorithm, receiving personalized recommendations and optimizing their decisions accordingly. But as she argues, optimizing agriculture in Africa is not primarily an information problem: It is a trust problem, a market access problem, a collective action problem and a financing problem — and addressing those challenges requires a human connection. She explores Kuza’s experiences offering an AI chatbot to Kenyan farmers, and shares lessons on how best to leverage both human and digital resources to support smallholders in Africa.