AI in farming

Twenty five years of farm records, put to work.

TWB Farms is among the first farms in Britain to run some of its planning, paperwork and buying through an AI agent built for agriculture. What has been done so far, and what Clive Bailye can offer others.

The story so far

Yield maps, soil tests, and now an agent that files the forms.


The farm has kept records for a long time: yield maps and soil tests going back twenty five years, field by field histories that reach back to Clive's grandfather, machinery costs, grain contracts. Like most farms it never had the hours to use them properly. AI changed that.

Clive began using large language models, and then agent based systems, in his own businesses as soon as they were good enough. They cut his spend on professional fees and software development and let him get a week's work done in a day. The obvious next step was an agent that understood a farm. He brought in a small team to build it with him, tested it on TWB Farms, and in 2026 set up Agigence to take it to market.

The agent, known to its first users as Giles, builds a private memory of a farm from its own records and from asking the farmer questions. It picks the most suitable and most economical AI model for each task rather than being tied to one supplier. A second system checks its work. The farmer approves every action, and nothing leaves a farm's account without the farmer's express permission.

Young crop through straw residue
Field by field, the agent knows the history

What it has done on this farm

The SFI application that nearly got missed.


When the second Sustainable Farming Incentive window opened in 2026, Clive was on his way to London. Weeks earlier the agent had worked through which options suited each field. On the morning itself it watched the government site, started the application the moment the window opened, flagged a discrepancy in one land parcel, and had the whole thing ready to review within half an hour. Clive checked and cleared it on his phone. It went in before midday. The national pot ran out that afternoon.

The same agent has analysed twenty five years of yield and soil data, suggested a new crop, found a local microbrewery looking for growers, and put a red diesel order out to tender and brought back the offers. Farmers Weekly told the full story in October 2026.

The tools are not ready for every farm yet, and farm data raises questions that have to be answered properly. That is why he thinks farmers should be in the room when they are.

What Clive offers

A farmer who has done it on his own farm.


Clive works with organisations that want to hear from someone who has actually used AI on a farm.

  • SpeakingConference talks and farmer meetings on what AI is already doing on a working farm, what it is not, and how to start. Plain English and his own numbers.
  • Board and advisory workHelping levy bodies, assurance schemes, government agencies, supply chain businesses and investors think clearly about AI, farm data and what farmers will and will not accept.
  • PolicyData ownership, assurance and traceability, the administrative burden on farms and how technology could cut it. He has briefed Defra, the Treasury and MPs from across the parties.
  • For farmersFarm Assist is in trials now, with Defra ADOPT funding covering the running costs for the first farms, and a wider release planned for 2027. Agigence has the details.

Where he stands

Progress, with the farmer in charge.

Clive's view is that AI is the next step in the same line as the tractor and the combine. It arrives with downsides, it changes jobs, and nobody who adopts it properly wants to go back. The risk is in how people use it. For farming the test is simple. Does it let farmers farm rather than fill in forms, does it keep their data theirs, and does it pay. If the answer is yes to all three, the industry should get on with it.

Talk to Clive about AI in farming

Speaking, advisory and press enquiries through the form. Choose the option that fits.

Get in touch