Skip to main content
lab-notes.aiPublic-service AI playbooks
Opportunity 05/17AgricultureEnvironment, land and water

Farm Advisory Support

Explore which farm measurements would be needed before a general recommendation could become field-specific advice.

Inspect starter dataSet up the build partner
Published sources
2
Starter data
18 synthetic records
Domain build partner
build-farm-advisory
Last reviewed

From the draft strategy

Opportunity

The draft strategy names AI for farming and precision agriculture as a potential public-service application, mentioning yield, livestock, water, and fertiliser use, without naming a decision a farmer would actually take with it or the evidence it would rest on.

Where the draft says it

Table 2 — potential public-service applications

Read the draft strategy source (opens in a new tab)

Source register

Research already done

Real published source

We reviewed these published sources so you do not have to start from zero. Each source shows what it covers, how it can be accessed, and why it matters.

2 sources

  1. Department of Agriculture, Environment and Rural Affairs

    Open

    Nutrient Management Plan

    What it covers
    What a nutrient management plan has to record field by field, including a valid soil nutrient analysis with phosphorus and pH, and the soil nitrogen supply index that fertiliser and manure decisions are set against.
    Why it matters here
    It defines the field-level measurements this example would depend on, which is why the synthetic records hold a soil pH and a nitrogen index and nothing about a holding or a person.

    Open the source (opens in a new tab)

  2. Department of Agriculture, Environment and Rural Affairs

    Open

    Agricultural Census in Northern Ireland

    What it covers
    The June census of Northern Ireland farms, published each year with crop areas, livestock numbers, and farm labour, alongside data tables and an interactive report; the 2025 results counted about 25,800 farms working roughly a million hectares.
    Why it matters here
    It is the published picture of what is grown and kept here, and it sits at farm counts and hectares — several steps above the single field any recommendation would actually be about.

    Open the source (opens in a new tab)

Starter data

Starter dataset

Synthetic working data

Use 18 synthetic rows across six field groups to explore an advisory question without using a real holding's records.

AI-assisted research helped identify and interpret the published sources. We then created a small, non-sensitive synthetic dataset shaped by the information those sources expose. It is for exploration—not evidence, training, or operational decisions.

Records
18
Fields
6

Field preview

table view
  • id

    SYN-FA-01 · SYN-FA-02 · SYN-FA-03

  • fieldGroup

    Field group A · Field group B · Field group C

  • crop

    Grass silage · Permanent pasture · Spring barley

  • soilPh

    6.2 · 5.8 · 5.5

  • nitrogenBand

    index 1 (moderate) · index 0 (low) · index 2 (high)

  • yieldBand

    10-12 t DM/ha · 8-10 t DM/ha · under 8 t DM/ha

Inspect all 18 recordsView repository file (opens in a new tab)Download JSON (opens in a new tab)

How it was prepared

AI authored fictional field groups with crops, soil pH, soil nitrogen supply indexes, and banded yields.

Limitations

  • The groups are letters and the figures are invented, so nothing here is a measurement of any soil, crop, or yield.
  • Six lettered field groups cannot stand in for the range of farm types, soils, and weather here, and there is no authoritative guidance table sitting behind the rows.
  • A pH and an index leave out the weather, the livestock, the costs, and the local knowledge that decide whether a suggestion is any use on the day.

Clone and build

Domain build partner

Build with a domain-aware coding agent

Clone the repository, then ask your coding agent to load the skill that ships with it. The skill brings the opportunity, sources, starter data, known unknowns and constraints into the conversation before anything is proposed. It follows the open Agent Skills standard, so it works in Claude Code, Codex and any other agent that reads skills.

01 / Clone

git clone https://github.com/Hypership-Software/lab-notes.ai.git

02 / Invoke

Claude Code
/build-farm-advisory
Codex
$build-farm-advisory

Conditions, not a checklist

Before you build

These are not footnotes. They are conditions that any responsible prototype would need to address.

  • Farms carry the cost of wrong advice

    Getting this wrong reaches livelihoods, animal welfare, water quality, and soil, and the farm carries that cost rather than whoever built the tool.

  • Digital records create representation gaps

    Farms with fuller digital records are better represented in the data behind any tool like this, so its advice tends to fit those farms best.

  • Improvement can move harm elsewhere

    Improving one input can quietly move the harm somewhere else — to water, emissions, biodiversity, or welfare — and no general model replaces walking the field or a qualified adviser.