Agro Controler

Agribusiness planning and management

WEB

Role: Senior Product DesignerPlatform: B2B web applicationScope: UX, UI, product structure, design system

A planning tool for farming companies that ties the season plan to real work in the field.
The team picks crops and fields, sees the money, checks the machines and hands out work orders.
I worked on the whole chain, not one feature: structure, heavy tables, calculations and visual hierarchy.

How a season works

Crop rotation
Economics
Machines
Plan
Bottlenecks
Field work

Problem

The season plan and the field work lived apart, so plans broke on contact with reality

A farm plans in spreadsheets: crops in one, money in another, machines in a third. The plan looks fine until it turns out there aren’t enough tractors in May.

You change a crop and have to recount the money by hand in another file.
Machine specs are there, but nobody can say how many hectares a set covers per day.
A shortage of machines shows up only when work is already late.
The plan ends as a report, and field tasks are made separately.

Research

I followed one season from the first decision to a tractor in the field

Mapped how a season decision flows: crop, area, money, machines, work.
Took apart the real calculations: yield, margin, machine output, fuel, time windows.
Looked for the moments where people stop and rebuild context by hand.

Insights

Each step changes the next

Crop drives money, operations drive load, machines limit the plan.

Numbers need a story

A spec table doesn’t say “we’ll be late in May”.

Lots of data isn’t the problem

Not telling state from guess from risk is.

A plan should end in work

Not in a report someone retypes into tasks.

Key questions

Q1
How do you let an agronomist change a crop and see the money change in the same place?
Crop rotation
Q2
How do you turn machine specs into a plain answer: hectares, fuel, cost?
Machines
Q3
How do you show the exact moment a plan stops being realistic?
Capacity
Q4
How do you turn the plan into tasks without losing why they exist?
Work orders

Decisions

Three rules for a screen full of numbers

DecisionInstead ofWhy

Edit right in the table, totals and charts stay beside it

Separate files and screens per step

You see the effect of a change where you made it

Split parameters, output and cost

One long spec table

“Which set is faster” and “what it costs” are different questions

Highlight shortages and risk, not every value

Equal emphasis on all data

Colour goes where you need to act

Details and fixes open from where the problem is

A separate optimisation section

Action stays next to its cause

Stable columns and clear grouping let people scan big tables without losing their place.

Design

One season, four screens

Q1 · Crop rotation

A crop choice and its money in one place

Set area, yield and price, and watch revenue and margin move.

Agro Controler

SettingsHelp & feedbackData synced 2 min ago

Northwind Farms · 18 863 ha / 2025 season

AAdmin

CulturesCrop rotationSeedsSZRPerformanceStaffProduction PlanOverheadEconomy

Crop Rotations

Total area18 863haAdd

Rotation 1Rotation 2Rotation 3

Crop / varietyArea, haYield, c/haPrice, USD/tGross margin, USD/haRevenue, USD/ha

Corn · Pioneer P92412 640782057791 599

Peas · Astronaute1 85028310178868

Sorghum · Milo 5057 38146165149759

Add

Financial • Rotation 111 871 ha of 18 863 haTotal revenue$11 429 339Average revenue$963/haGross margin$3 485 629Average gross margin$294/ha

Crops outside rotationCrop / varietyArea, haYield, c/haPrice, USD/tGross margin, USD/haRevenue, USD/ha

Sunflower · P64LE252 42024410224984

Winter wheat · Skagen2 872521983801 030

Soybean · ES Mentor1 700274204041 134

Add

Financial • Crops outside rotation6 992 ha of 18 863 haTotal revenue$7 267 240Average revenue$1 039/haGross margin$2 320 240Average gross margin$332/ha

Area distribution by crop

22% Corn · North Field16% Peas · Central Field62% Sorghum · Steppe Block

Gross margin by crop, USD/ha

10008006004002000

Corn

Peas

Sorghum

Crop rotation planning with economic indicators and chartsOpen full size ↗

Q2 · Machines

Specs turn into hectares, fuel and cost

Compare machine sets by what they actually get done.

Comparison of equipment performance, cost, and configuration parametersOpen full size ↗

Q3 · Capacity

You see when the plan stops being real

Needed vs available machines through the season, with the gap and the work at risk.

Agro Controler

SettingsHelp & feedbackData synced 2 min ago

Northwind Farms · 18 863 ha / 2025 season

AAdmin

CulturesCrop rotationSeedsSZRPerformanceStaffProduction PlanOverheadEconomy

Production Plan

CornSunflowerWheatRapeseed

Add process

12.08.2024Pre-seasonSoil preparationMain operationStubble disking2 640 ha0.180 machine h/ha8.6 l/ha

26.08.2024Pre-seasonCrop careMain operationFertilizer application2 640 ha0.085 machine h/ha2.3 l/ha

26.08.2024Pre-seasonCrop careAuxiliary operationFertilizer transport2 640 ha0.040 machine h/ha1.6 l/ha

16.09.2024Pre-seasonSoil preparationMain operationDisking2 640 ha0.160 machine h/ha7.9 l/ha

30.09.2024Pre-seasonSoil preparationMain operationDeep ploughing2 640 ha0.420 machine h/ha21.4 l/ha

27.03.2025Pre-sowingSoil preparationMain operationEarly harrowing2 640 ha0.070 machine h/ha3.2 l/ha

10.04.2025Pre-sowingSoil preparationMain operationPre-sowing cultivation2 640 ha0.150 machine h/ha7.4 l/ha

22.04.2025SowingSowingMain operationFertilizer-assisted sowing2 640 ha0.190 machine h/ha5.8 l/ha

22.04.2025SowingSowingAuxiliary operationSeed transport2 640 ha0.020 machine h/ha0.9 l/ha

22.04.2025SowingSowingAuxiliary operationSeed loading2 640 ha0.015 machine h/ha0.4 l/ha

06.05.2025VegetationCrop careMain operationHerbicide application2 640 ha0.055 machine h/ha1.4 l/ha

06.05.2025VegetationCrop careAuxiliary operationWater transport2 640 ha0.030 machine h/ha1.1 l/ha

04.06.2025VegetationCrop careMain operationInter-row cultivation2 640 ha0.120 machine h/ha4.9 l/ha

02.10.2025HarvestHarvestMain operationDirect combining2 640 ha0.310 machine h/ha16.5 l/ha

02.10.2025HarvestHarvestAuxiliary operationGrain transport2 640 ha0.140 machine h/ha6.2 l/ha

Equipment capacity shortage inside the production planOpen full size ↗

Q4 · Work orders

The plan ends in field work

Dates, operation, crop, materials and a person responsible, per field.

Field tasks and active work ordersOpen full size ↗

Result

Success is a plan that actually gets done

The real question: does the tool help build a realistic plan, catch bottlenecks in time and turn decisions into finished work?
So we measure the chain itself, step by step, instead of screen views.

01Plan created
02Risk found
03Decision made
04Task created
05Work done

Design system

One visual language for planning, calculation and field work

Agro Controler connects planning, calculations and daily execution, so the design system has to work equally well for dense tables, analytical charts and short operational actions. Foundations set the shared visual language, semantic tokens separate an element’s role from its exact colour, and product patterns keep the same logic across modules.

The system is built in three layers: primitive values measured from the product, semantic roles used by every screen, and component tokens for decisions that belong to one component.

Neutral surfacesSix dark steps carry the whole interface hierarchy. Elevation is a lighter surface, never a shadow.

neutral.950#0d0d0eApp background
neutral.900#1b1b1cWorkspace panel
neutral.875#212122Row stripe, active nav
neutral.850#252526Card, summary
neutral.800#2a2a2bSide panel
neutral.750#323233Selected tab, button

Text & border alphas

white.100
white.70
white.50
white.30
white.10
white.05

Text and dividers are white at fixed opacities, so they stay consistent on every surface.

Semantic tokensScreens use roles, not hex values.

Background & surface

color.background.appneutral.950
color.surface.primaryneutral.900
color.surface.secondaryneutral.850
color.surface.elevatedneutral.800
color.surface.selectedneutral.750

Text

color.text.primarywhite.100
color.text.secondarywhite.70
color.text.mutedwhite.50
color.text.disabledwhite.30

Border

color.border.subtlewhite.05
color.border.defaultwhite.10
color.border.focusblue.400

Status

color.status.successgreen.300
color.status.warningsand.200
color.status.criticalred.300
color.status.infoslate.400

Data visualisationCategories and thresholds never share a colour logic: a crop colour cannot be read as a warning.

Crop categories

Corn#518a4d
Sorghum#edd25b
Rapeseed#d98a3d
Barley#7fb8a4
Chickpea#c7a878
Sunflower#6bb1dd
Wheat#ae4da3
Soybean#9e5050

Capacity threshold

Available capacity
Required capacity
Shortage

Process series

Disking
Cultivation
Crop protection

Conclusions

In complex B2B, the design is the chain, not a screen

A money assumption can be checked against machines, turned into a decision and sent to the field.
Dense data stays readable because hierarchy comes before decoration.
I worked at the level of the whole system: data, calculations, flows and the design system.