IoT-driven precision fertilizer management platform for large-scale arable farming operations.
We built a field sensor network, edge processing layer and agronomist dashboard that replaced blanket fertilizer application with real-time, field-specific recommendations across a multi-thousand hectare operation.
- NVZCompliance fully documented end-to-end
At a glance
- ClientA large-scale arable operation with a seven-figure annual fertilizer budget.
- ChallengeDecision intelligence in agronomy: every input has cost, environmental, and regulatory implications.
- ApproachThe complete fertilizer decision loop: measure, analyse, recommend, apply, verify.
- ResultsFrom blanket application to field-level intelligence.
Fertilizer was applied on a fixed schedule. The cost was real, the waste was structural.
The client grows cereals and oilseeds on varied soils. Fertilizer followed growth stage and historical averages: same rate, every field, same date, whatever the soil held.
Fields with enough residual nitrogen got the same dose as depleted ones, and leaching from over-application became a regulatory and environmental liability as nitrate vulnerable zone (NVZ) rules tightened.
Sensor data without agronomic intelligence is noise.
From zero, with no prototype, the platform had to both measure the soil and turn raw readings into field-level recommendations a farm manager could act on.
Sensor network at field scale
Dense coverage of hundreds of fields, in variable weather and minimal connectivity, forced hardware and edge-processing decisions standard IoT deployments don’t face.
Agronomically valid recommendations
Soil NPK readings alone aren’t a recommendation: the engine also needed crop type, growth stage, weather forecast and regulatory limits.
Spreader integration for closed-loop application
The prescription map had to leave the dashboard and reach the spreader in a format it could execute autonomously, zone by zone.
Connectivity in low-infrastructure environments
Many fields had no mobile signal or WiFi, so the edge layer had to buffer data, run recommendations offline and sync later.
From soil sensor to spreader prescription. End-to-end.
Each component owns one stage of the loop, and data flows between them continuously, so recommendations improve as yield and application history accumulate.
Inputs
- Soil sensors
- Crop & growth context
- Weather & regulatory limits
Decision engine
AI dosage engine
Outputs
- Agronomist dashboard
- ISOXML spreader prescription
- NMP compliance report
- 01Sensor layer
IoT soil sensor network with edge processing
Multi-parameter nodes measure moisture, temperature, electrical conductivity, pH and ambient conditions, sending over LoRaWAN to edge gateways that buffer through connectivity gaps.
- 02Field intelligence
Field zone mapping and soil variability modelling
Sensors, historical soil surveys, satellite NDVI and elevation models build each field’s variability map; zone boundaries update as new readings arrive.
- 03AI dosage engine
Nutrient recommendation engine with crop and regulatory context
Sensor readings plus crop type, growth stage, yield history, weather forecast, soil texture and NVZ limits produce a variable-rate prescription per zone.
- 04Agronomist dashboard
Web dashboard with field map, sensor status, and prescription review
Fields are colour-coded by nutrient status; agronomists review, adjust and approve each prescription before release, seeing its economic and environmental impact.
- 05Application
Variable-rate spreader integration and application logging
Approved prescriptions export as ISOXML, the spreader adjusts rate automatically zone by zone, and GPS-tracked application data comes back after each run.
- 06Compliance & reporting
Nutrient management plan documentation and regulatory reporting
Every application is logged — field, date, product, rate, GPS coverage — and 5 years of history serve farm assurance audits and lender sustainability reporting.
Nine capabilities spanning IoT, AI, dashboard, and compliance.
- IoT soil sensor network
- Field zone mapping
- AI dosage engine
- Agronomist dashboard
- Spreader integration (ISOXML)
- Mobile app for field staff
- Satellite NDVI integration
- Nutrient Management Plan reports
- NVZ compliance alerts
Cost down 23%. Yield up 12%. Compliance fully documented.
Fertilizer cost down 23% in first season
Ending over-application in well-stocked zones delivered the largest single input saving in the operation’s history.
Yield up 12% per hectare
Targeted dosing of previously under-nourished zones drove the gain, particularly in variable-texture fields where rates varied most.
850+ sensor nodes active
The network covers the whole operation, dense enough to capture within-field variation that justifies different prescriptions per zone.
NVZ compliance fully documented
Every decision is logged automatically and the Nutrient Management Plan builds itself from application records, removing weeks of manual admin each season.
From the team
The AI dosage engine did not replace agronomist judgement — it structured it. Clear-cut zones are prescribed automatically, with confidence intervals that reflect sensor coverage density. Zones with thin coverage are flagged for review. The system says where it is uncertain; the agronomist decides whether to accept or inspect.
Stack
Offline-first and built for sensor scale: every layer has a track record in field deployments without guaranteed connectivity.
Building a precision agriculture platform?
We’ve shipped sensor networks, AI recommendation engines and precision application integrations for large-scale operations. It’s worth a chat.
More case studies
