KARAZ · Knowledge for AI-Driven Resilient Agro-Zones

The calendar is a memory of climates that no longer happen.

Human + Machine, for crops you plant once.

And the market has stopped waiting too. A grower of perennial crops commits for decades against two things that now change every few seasons.

You cannot replant your way out of that gap. You can only decide your way through it.

What KARAZ is See what arrives
Three clocks, one decision
The tree you plantProductive lifetime 20–30 yr
The climate you planted it forBefore the assumptions shift ~5 yr
The market you planted it forBefore demand and price move 1–2 yr
this gap is the whole problem
What KARAZ is

Three questions a grower cannot answer alone.

KARAZ reads the field and the market together, and answers at three time horizons. Most agricultural software answers the first. Some answer the second. The third is the one that decides whether an orchard is still viable in 2040.

This week What do I do now?

Irrigation, fertigation, protection and intervention — issued against measured stress in the block, not against a schedule written for a climate that has moved.

This season When do I pick, and who buys it?

Predicted arrival and quality per block, matched to the market window where the premium actually is — before pickers, buyers or contracts are committed.

This decade What should I be growing?

Which varieties, at which altitude, in which zone — the twenty-year decision, taken against measured evidence instead of the way it has always been done.

The engine · Human + Machine

The first feed is not a sensor.

Your grower knows why last March went wrong. No model has that, and no amount of data replaces it. It goes in first, alongside four machine feeds — and what comes out is what they mean together, which is never what any of them says alone.

Grower knowledge — why last season went wrong
nobody else starts here
In-field sensing — soil moisture at depth, canopy temperature, microclimate, stem diameter
Satellite — vegetation and water indices at 10 m, every few days
Climate — reanalysis and forecast, at block resolution
Market — arrival windows, prices, demand segments
Output 01Irrigation & fertigation
Output 02Protection & intervention
Output 03Harvest timing
Output 04Market window & placement

Everything happens pre-harvest. Once the fruit is off the tree it is logistics, not agronomy. KARAZ works in the weeks where the outcome can still be changed — which is also the only window where advice is worth paying for.

Not a monitoring platform. Monitoring tells you what happened. KARAZ links the agronomic decision to the market result, which is the loop nobody else closes in perennial crops — and it runs on the trees already in the ground, not on a variety you would have to plant and wait a decade for.

Inside the engine
Two crops, two jobs

One loop. Different logic per crop.

The sensing, the pipeline and the data architecture are shared. The quality model and the market logic are not — because an orchard and a vineyard fail in different ways.

Cherry & stone fruit · interventionist

Measurable control over quality and timing.

Six quality parameters are still steerable in the last six weeks. The seventh variable — the date the fruit arrives — decides what the other six are worth. KARAZ steers all seven.

  • Steer the fruit. Size, sugar, colour, firmness, cracking resistance and shelf life, against measured stress rather than a calendar.
  • Protect the window. Frost, heat and pest pressure flagged with the date by which acting still helps.
  • Avoid the glut. Predicted arrival routed to where the premium is, block by block.

Control range · harvest timing · size · brix · firmness · colour · cracking · shelf life

The cherry branch →
Vine · interpretive

The right wine for the right market.

A vineyard decision does not change a number on a screen. It changes sugar and acid, which changes style, which changes who buys it. Most estates learn which wine they made a year after they could still have changed it.

  • Read the vintage early. Water deficit, canopy stress and veraison tracked live, so the style is known before it is fixed.
  • Steer it while you can. Irrigation is new to Mediterranean vineyards — a structural input nobody has fifty years of instinct for.
  • Match it to a buyer. The resulting profile mapped to the demand segment that pays for it, reducing what goes unsold.

Targets · water ≥20% · pesticide ≥30% · yield ±7% · sell-through +10pp · embedded emissions ~10%

The vine branch →

Olive follows the same pattern — its own quality model and market logic on the shared engine. Roughly 60–70% of the infrastructure carries to each new crop.

What actually arrives

Decisions, not dashboards.

A dashboard shows a number and leaves the thinking to you. KARAZ issues a decision: which block, what to do, why, and the date after which doing it no longer helps. If there is nothing worth acting on, it says nothing.

Critical BLOCK 12-CCherry · 1,250 m
Cover the early block before Thursday night. Forecast dips to −1.4 °C at 04:00 with buds at first white. Exposure at this stage historically costs a third of the block.
Act beforeThu 18:00 Thenno longer useful
Pending BLOCK 07-ACherry · 1,350 m
Hold the pick three days. Sugar is four days behind colour. The fruit looks ready and is not. Picking on colour alone costs Brix at the grade that pays, and the window is not closing yet.
Re-checkIn 72 h Windowstill open
Scheduled PARCEL 04Vine · 400 m
Two irrigation events, 18 mm total, before 24 July. Veraison six days early and deficit building. Left alone, this vintage lands riper and heavier than the estate's buyers pay for.
Act before24 Jul Moves pick to8–14 Sep
Market ESTATECherry · all blocks
Contract the late block forward now, not in June. Predicted arrival falls inside the regional glut. Buyers still have room today; in five weeks they will not, and the same fruit clears at a fraction.
Act before30 Apr Exposure62% of crop

Decision cards illustrate the format and the severity model. Blocks and zones are coded.

The parent

Born where the smallest margins decide who survives.

Huggin Munin has been building data-driven supply-chain systems since 2014 — in economies where data was not a strategy, it was the only competitive advantage available. What we learned moving energy now moves produce.

The group and the method →
FleetserveEnergy Agricom · KARAZAgriculture EasyOrderFMCG

Huginn is thought — he flies out and collects the world. Muninn is memory — he mines what was gathered and says what it means. Data collection and data mining, named a thousand years before either had a name.

Let's talk

If you plant trees, we should talk.

Tell us your crop, your altitude, and what your last bad season cost you — even if you are not sure what your next step is. No pitch, no strings attached. One call, and we will tell you within thirty minutes whether we are the wrong people for it.

Start the conversation Read the evidence