LOCAL FOOD / AI BEYOND TECH
Farm Run
The farm does not need to become an AI company.
Local food already has farms, farm stands, Food Hubs, grocers, and ways to buy. What is often missing is a useful bridge between what happens to be available this week and what a household knows how to do with it.
Farm Run explores whether AI can close that gap without replacing the relationships and buying channels that already work.
The difficult data is a sentence.
Traditional software wants clean inventory, stable categories, exact counts, and an API. The reality of a small farm may sound more like: the fennel is especially good this week, the tomatoes are nearly ready, and there is more chard than expected.
That information is irregular, qualitative, and time-sensitive. Historically, turning it into structured customer guidance required more time than a small producer could reasonably give. Language models create a new possibility: software that can begin with ordinary descriptions and work with their meaning before demanding a perfect data system.
Turn this week’s reality into a reason to buy.
A short farm update could become seasonal recipes, meal ideas, an editable shopping plan, or a timely customer message. Produce that might otherwise be unfamiliar becomes something a household can picture cooking and knows where to find.
The purpose is not to generate more generic food content. It is to translate a specific local reality into useful action, then send the customer back to the farm stand, Food Hub, or grocer that already owns the transaction.
Add intelligence, not another intermediary.
Technology companies often respond to a fragmented market by building a marketplace and placing themselves in the middle. Farm Run takes the opposite position.
It does not need to own inventory, checkout, payment, ordering, or fulfillment. It can remain a thin demand layer: helping farms express what matters, helping customers make plans, and helping the existing network coordinate without extracting the relationship into another platform.
A small loop points to a much larger opportunity.
Once a week, a producer describes what has changed. AI develops useful possibilities from those facts. A person reviews anything that will reach customers. Shoppers receive guidance they can act on through the channels they already trust.
The current prototype demonstrates that loop through supply updates, customer planning, human review, and handoff to existing sellers. It does not yet use a live model or live partner data. The next proof is whether one farm or Food Hub can run the weekly cycle with little enough effort to change real customer behavior.
That question reaches beyond agriculture. Large parts of the economy are not waiting for another enterprise platform. They are waiting for a low-burden way to turn messy, local knowledge into coordinated action. Farm Run is one concrete test of what AI could contribute there.
Owen Fowler
AI Systems Builder
Tell me what you’re working on.
A few sentences is enough.