On a warm morning in Lurambi, Kakamega County, Peter Wekesa presses a metal probe into his maize farm and waits for a reading.
It offers a simple explanation for a problem that has puzzled him for years: declining yields on land he believed was still fertile.
For years, he has followed a script familiar to millions of Kenyan farmers: apply DAP at planting, top-dress with CAN, pray for rain and hope for the best.
Yet season after season, his yields have stagnated, even as fertiliser prices soar. “I have done everything right,” he says. “But the harvest keeps getting smaller.”
Wekesa’s experience mirrors a national crisis quietly unfolding across Kenya’s farmlands.
Studies by the Kenya Agricultural and Livestock Research Organisation (KALRO) and the Ministry of Agriculture show that more than half of Kenya’s arable land is degraded, with widespread soil acidity, declining organic matter and severe nutrient imbalances.
In high-potential regions such as Western Kenya, years of monocropping particularly maize and sugarcane, have stripped soils of key micronutrients, leaving farmers trapped in a costly cycle of low productivity.
Despite successive government interventions, from fertiliser subsidy programmes to liming initiatives meant to correct soil acidity, yields have remained stubbornly low in many areas.
Soil scientists say the problem is not only what farmers apply on their fields but also the lack of precise information guiding those decisions.
“Most farmers are managing soil without really knowing what it needs,” says Njuguna Mwaura, team lead at Angaza, a Kakamega-based technology firm piloting a mobile application that integrates rapid soil testing, artificial intelligence and real-time farm advisory services.
“Farmers apply fertiliser before testing their soil. That is blind management,” he says. Angaza’s platform combines handheld soil sensors with an AI-powered mobile app. Once inserted into the soil, the sensors measure up to eight parameters — including pH, temperature, nitrogen, phosphorus, potassium, salinity and electrical conductivity — and transmit the data to the app.
Within minutes, the system generates tailored recommendations, advising farmers on what nutrients are deficient, what inputs to apply and what to avoid.
“This allows farmers to move away from blanket use of CAN and DAP,” Mwaura explains. “If your soil lacks potassium, adding more nitrogen will not solve the problem. You only spend more money and worsen soil health,” he adds.
Preliminary findings from Angaza’s pilot in Kakamega County underscore the scale of the challenge. According to the firm, most farms tested so far have highly acidic soils with low levels of essential nutrients, particularly potassium.
“Many farmers are shocked when they see the results,” Mwaura says. “They realise they have been applying the wrong fertiliser for years.”
National data supports these observations. The Ministry of Agriculture estimates that soil acidity affects over 30 per cent of Kenya’s arable land, especially in the Western and Central regions.
Acidic soils limit nutrient uptake, suppress crop growth and significantly reduce yields, even under favourable rainfall conditions.
In response, Kenya has rolled out multiple soil health interventions, including donor-funded programmes aimed at restoring degraded soils and promoting sustainable land management.
However, adoption has been slowed by limited access to soil testing services and overstretched extension systems.
Angaza believes digital tools could help bridge that gap. Beyond soil analysis, the app allows farmers to photograph crops showing signs of disease or pest damage. The AI analyses the images and suggests corrective measures, effectively turning a smartphone into a diagnostic tool.
“It’s also designed to support extension officers. Instead of relying only on book knowledge, they can give farmers real-time, data-driven advice while standing on the farm,” Mwaura says.
The platform integrates short- and medium-term weather forecasts, offering 30-day and 120-day projections. More importantly, it translates forecasts into practical guidance.
“It doesn’t just say it will rain or not rain. It tells the farmer what action to take based on their crop and location,” says David Mandala, one of the developers behind the system.
Mandala says the idea of the app grew out of frustration with what he describes as expensive but unreliable agricultural data.
“We saw farmers spending heavily on inputs, yet the information guiding those decisions was often inaccurate or too general,” he says.
To address this, the system runs multiple analyses every minute after data is transmitted from the soil sensor, continuously validating results before issuing recommendations.
The developers estimate the platform’s accuracy at about 92 per cent, a figure they say improves with repeated sampling over time.
In addition to soil and crop diagnostics, the app incorporates satellite mapping to help farmers better understand variations across their land.
Angaza is currently at the minimum viable product stage, piloting the technology with the Kakamega County government and the Bukura Agricultural Training Centre. The firm expects to roll out the product more widely within the next two months, starting with county agricultural extension officers.
County officials, grappling with declining productivity in a region once considered Kenya’s breadbasket, have expressed interest in the approach.
“Our ambition is to roll this out nationally and eventually beyond Kenya,” Mwaura says. But scaling up will require more than functional software. Mandala says expansion will depend on securing funding, building partnerships with county governments and development agencies, and investing heavily in farmer training.
Beyond the app itself, the model requires soil sensors, smartphones, reliable internet connectivity and sustained extension support, resources that remain scarce for many smallholder farmers.
“Teaching farmers to change how they farm is the hardest part,” he says. “You can’t just deploy an app and walk away.”
Angaza is therefore seeking partnerships that would embed the technology within existing public extension systems rather than operate it as a stand-alone product.
For farmers like Wekesa, seeing his soil’s condition translated into clear, actionable data has already changed how he farms.
“At least now I know what is wrong and that makes a big difference,” he says.
As climate change tightens its grip, fertiliser prices rise and Kenya’s soils continue to degrade; the future of food security may depend less on how much farmers apply to their land and more on how well they understand it.
By Chris Mahandara
