For banks: lend on the field, not on the average
Before the loan, FairGrow shows what this plot has actually done over the past seasons. After the loan, sensors, crop records and an API keep you informed about the crop you financed. Satellite, weather and terrain, with nothing installed in the field.
Two moments where a field tells you something
A loan application names a farm, a crop and an expected yield. What it does not show is the field itself: whether it was farmed last season, where the cold air collects on a frost night, whether the water reaches the roots in August. FairGrow puts that on the table before the decision, and keeps it coming in after the money has left.
Does the plan hold up on this plot?
Within a few days of receiving the locations you get a report for every field in the application, built from Copernicus Sentinel-2 imagery, the Copernicus elevation model and the ECMWF and DWD weather models.
Is this land actually farmed?
The vegetation history runs back to 2017. A plot that has been green every season looks different from one that was abandoned two years ago, and the curve shows which of the two you are financing.
Does the crop fit this field?
Elevation, slope, aspect and modelled soil moisture by depth tell you whether an orchard here needs irrigation to survive a normal August, and whether the water balance shows that it got it.
Where the cold air collects
On a clear night cold air flows downhill and settles in low, enclosed plots. Terrain, the elevation profile and measured inversions show which orchard in a village loses its blossom and which one, a few metres higher, does not.
Fifteen years of frost dates
The last spring frost and the first autumn frost for this location, year by year, including how often the last frost landed inside the blossom period. That is the difference between a climate class and a date you can plan around.
Water: rain against crop use
Rainfall against reference evapotranspiration gives a water balance per week, with modelled soil moisture by depth beside it. A shower that wets the top centimetre and never reaches the root zone shows up as what it is.
The neighbours as a benchmark
Plots in the same village share the weather and little else. Comparing the financed field with the fields around it turns one number into a position: above or below what this season allowed here.
A regional average hides the borrower
On one Sentinel-2 scene, measured on the same day under the same weather, the vegetation index ran like this across five points inside a single village.
- Irrigated field, 500 m south: 0.65
- The village itself: 0.54
- The financed orchard: 0.45
- Neighbouring plot, 300 m north: 0.24
- Bare slope, 1 km north: 0.11
A regional average cannot separate those borrowers. The elevation difference between the highest and the lowest of these points is 108 metres, which on a frost night is the whole story.
Follow the crop, not only the repayment schedule
A loan runs for years. Most of what goes wrong in that time is visible on the plot long before it reaches an instalment.
Measured instead of modelled
FairGrow sensor nodes measure soil moisture, soil temperature, air temperature, humidity and light every few minutes and send their readings through a gateway on the farm, over HTTPS with a token per device, buffered when the connection drops.
- Timestamped data for the plot itself, not for a grid cell of several kilometres
- Optional: the assessment and the weekly monitoring both work without them
- Currently in the certification process
- Quoted separately, for the plots where you want measured evidence
What the farmer actually did on the field
The grower registers the crop and variety, sowing or planting, irrigation, crop protection, fertilisation and harvest in the crop records, because a spraying file is becoming mandatory anyway. With explicit consent from the farm, per type of data and revocable at any time, those records and the sensor and satellite series are readable through the partner interface.
- Operations with date, plot, product and dose, checked against the national product register
- Irrigation and crop water use per plot, from the irrigation advice
- Sensor series and satellite indices for the same plot, in the same call
- Corrections keep their version history, so a record cannot be quietly rewritten
The partner interface is built and is released with pilot partners, so your own monitoring systems can hold the field data next to the loan file.
Three things that arrive without being asked for
A report per plot
Weather, satellite, water balance, night inversions, frost outlook and terrain, in an eleven page PDF to up to five addresses, for example the farmer and the relationship manager. An example report for a fruit orchard is ready to send.
The farmer first
Twenty-five alert types, checked against the forecast every thirty minutes, by e-mail and Telegram. Frost always comes through, even during quiet hours, because that is the night a crop can still be saved.
Aggregated to you
How many financed plots carry a high frost signal tonight, by village and by borrower, so relationship managers can follow up. This view is built together with the first pilot partner.
What this lets you answer
| Your question | What FairGrow shows | When |
|---|---|---|
| Is this plot farmed, and how well? | Vegetation history per season since 2017, next to neighbouring plots | Before the decision |
| Does this field carry frost, drought or flood risk? | Frost dates over fifteen years, cold air pooling from terrain, water balance, water flow paths | Before the decision |
| Is the crop developing normally this season? | Vegetation and moisture indices at 10 m, updated every few days | Weekly, during the term |
| What did the farmer actually do on the field? | Crop records: planting, irrigation, treatments, harvest, with consent | Continuously, by API |
| What were the conditions on the night of the damage? | Hourly weather record, alert log, imagery before and after the event | After the event |
| Which plots in my portfolio are at risk this week? | Aggregated signal per village and per borrower | Weekly |
The file exists before the claim is filed
When frost, hail or drought strikes, the local weather record of that plot, the alert log, the satellite scenes before and after the event and the terrain context were all recorded beforehand. Nobody has to reconstruct a night afterwards from memory.
- Every alert is logged: what the farmer was warned about, and when
- Satellite scenes are dated and cloud filtered, so a drop in vegetation has a date
- Thunderstorm conditions and gusts for the hour of the damage
- A claim file export for insurers is pilot development, not a finished product
What FairGrow does not do
FairGrow does not score credit, value collateral or make a lending decision. It supplies evidence about a field. The assessment stays with you, and so does the responsibility for it.
Without sensors, night temperature, wind and soil moisture are modelled rather than measured: the weather model sees a grid cell of 2 to 11 km, and the differences inside that cell come from terrain and satellite. Hail, mudflows and landslides are forecast as weather conditions, not as events. Flood and landslide risk is limited to rainfall intensity, slope and wetness, with no dedicated model behind it. A clouded fortnight leaves a gap in the satellite record. You are better off knowing that before the first report than after it.
Where a national climate risk map tells you which region is exposed, this tells you which field is. The two levels do not compete: screening on the map, verification and monitoring on the plot.
The farm owns the data, you get access to it
A monitoring tool that only serves the lender gets switched off by the farmer within a month. FairGrow is built for the grower first: better timed irrigation, frost protection on the right night, crop protection when the models say it matters, in 57 languages on a phone. The better those decisions, the smaller the chance that a weather event turns into a default.
Certified and European
EasyData B.V. is ISO 27001 certified, certificate NL/1543IS/1618. Processing runs on our own servers in a Dutch data centre, under the GDPR, with demo and customer data strictly separated.
Per type of data
Nothing reaches a lender without explicit consent from the farm, given per type of data and revocable at any time. Access is granted per user and per device.
The file is not a hostage
Field data belongs to the farm, and it can always be exported in a form you can hand over. A file you can only read inside one platform is not a file but a hostage.
Eight weeks, five plots, no obligation
The assessment needs only satellite, weather and terrain data, so a pilot costs you nothing but the time to choose the plots. It is designed to answer one question: does field level information improve your assessment and monitoring compared with a regional average alone?
1. Send the plots
Five to twenty locations: coordinates, cadastral codes or outlines. No loan data and no client data are needed, and the satellite assessment itself needs no consent, because these are public sources about a location.
2. Look back at the season
Vegetation history, frost nights, heat and drought per plot, next to what you already know about these farms. Ideally you pick a few plots with a known outcome in 2026, including the ones that went wrong.
3. Follow them live
Weekly reports and alerts for eight weeks, with the first version of the portfolio view, and an optional extension through the spring frost period of March to May 2027. At week eight we evaluate together and you decide whether anything follows.
What lenders ask us
Do we need the farmer’s permission to start?
For the plot assessment from satellite, weather and terrain data, no: those are public data sources about a location. For sensor readings and crop records, yes, always, per type of data, and the farm can withdraw that consent at any time.
How far back does the history go?
To 2017 for Sentinel-2 satellite imagery, and to 2011 for the weather and frost record of a location, taken from the ERA5 reanalysis. Terrain does not change from year to year and comes from the Copernicus elevation model at 30 m.
Which countries does this work in?
Anywhere. The satellite, terrain and weather sources cover the whole world and the platform runs in 57 languages, so a plot in another country needs no different setup. Only the national product register behind the crop protection check is country specific.
Can we get the data into our own systems?
Yes, through the partner interface, per farm and with consent. It is built and is released with pilot partners. Plot level exports are available in the meantime.
Is this a replacement for a field visit or an agronomist?
No. It tells you where a visit is worth making and what to look at when you get there, and it gives you a record for the plots nobody will ever visit. Values are indicative and meant for early risk assessment, not as a substitute for an agronomic or insurance judgement.
What does it cost after a pilot?
Terms for an institutional programme depend on the number of plots and are agreed separately. For individual growers the platform runs from the Starter plan at no cost up to Estate at €39.50 per month; see the pricing page. Sensors are hardware and are quoted separately.
Start with five plots from your own portfolio
Send the locations and we will show you what FairGrow sees where a map shows only the village. Without obligation, and with nothing installed in the field.
Send us five plots → Open an account first →