wajenzi.

Measured in the field

A soil sensor on the plot, radar from orbit and ECMWF forecasts, fused into one answer: plant now, or wait.

See how it works

Three layers of data

The probe knows the plot, the satellite sees the district, and the forecast sees the week ahead. No one source is enough on its own.

Sentinel-1 radar

Satellite radar

Reads surface moisture and flooding through monsoon cloud and Harmattan dust.

15%water to plant maize

ReadingThreshold

Soil sensor

Moisture and conductivity, read at the root zone of the plot itself.

Weather forecast

ECMWF’s medium-range model, corrected with readings from the ground.

93%

rainfall accuracy shown by fused forecast systems in West Africa

Built for the field, not the lab

Every reading must end in one action a farmer can take, on a phone they already own.

More about us
PrototypeMini-prototype in bench testing
2GAny basic phone works

Every plot on one screen

What the engine sees before it sends advice. Shown here with sample data from twelve plots. See real readings from our test logger.

Sample · 12 plotsPlot 4 · PlantPlot 7 · 11% waterPlot 9 · WaitSoil moisture: low to high

Ready to plant

7/12

Plots above 15% water

Node battery82%

Messages heard91%

Soil temperature

At 10 cm, Plot 4

27°C

Today

Plot 4 crossed 15% water06:40

Sentinel-1 pass, radar updated09:15

Plant advice sent to 7 farmers13:02

Rainfall · 7 days31 mm

Mon
Tue
Wed
Thu
Fri
Sat
Sun

Forecast

27°Mon

28°Tue

25°Wed

21°Thu

24°Fri

One plot, five signals

15%

Water by volume at the root zone. Maize needs about this much to establish; on sandy loam we look for 15–20%.

From soil to phone

Measure

Sensors on the plot read soil moisture and nutrients at the root zone.

Combine

Radar and ECMWF forecasts are fused with the readings from the ground.

Decide

The engine checks the soil against what the crop needs to germinate.

Deliver

One instruction goes out by USSD or voice, in the farmer’s language.

The numbers we design around

32–99%

Maize yield gain from microdosing in Ghana trials

71%

Leaf area maize loses to early drought

56 dB

Extra radio loss through a tropical canopy

€55k

Five-year cost of a 100-sensor network

The ground corrects the model

ECMWF’s land model estimates soil moisture from rain, sun and evaporation, but Africa has few stations to check it against. A Kalman filter weighs each reading from our sensors against that estimate, pulling the forecast back toward what is really in the soil.

AdviceOne instruction per plot
Soil sensor
Satellite
Forecast
Phone
Northern Ghana

The question we answer

Is the soil wet enough to plant maize today?

Asked in Dagbani, answered in Dagbani

By USSD or a voice call, on any basic phone.