A four-county Michigan pilot launched in July aims to rewrite how federal crop insurance prices risk, shifting premiums from county-level averages to field-by-field rates built on satellite imagery, soil science and farm management records.

What does the pilot change?

The Michigan Department of Agriculture and Rural Development partnered with Aon, The Nature Conservancy, the Resilience in Agriculture Lending and Insurance Coalition, and a startup called SAGE to design the program for corn growers. The partners will layer satellite imagery, terrain data, soil science, weather history and farm-management records to identify what has been grown where, under what conditions, and how individual fields have performed over time.

Why county rates miss the real risk

USDA's crop insurance program covers more than 130 commodities and several hundred million acres. Premiums are typically set by crop type at the county or regional level, and USDA subsidizes a large share of every farmer's premium. Two farms in the same county with different soils, topography and management often pay similar rates, even when their risk profiles diverge sharply.

In their op-ed, former USDA Undersecretary for Farm Production and Conservation Robert Bonnie — a SAGE co-founder — and Aon Agriculture Resilience Center Senior Vice President Jonathan Sherrill put it bluntly: "Risk is determined field by field. But the farmer's price for USDA crop insurance isn't."

Why RMA hasn't moved yet

USDA's Risk Management Agency has explored proposals that would reward growers for soil-health practices such as no-till, cover crops and managed grazing. RMA has not advanced them because the data and analytics needed to set actuarially sound, field-specific rates have not existed. Pricing a single field differently requires knowing its soil type, topography, cropping history, and how management will change its risk going forward, for every field in the country.

What AI changes

Recent advances in artificial intelligence, agronomic modeling and field-level data integration are beginning to close that gap. A growing body of agronomic research shows that practices improving soil health and function can reduce drought and weather-related losses. Healthier soils often mean lower risk. The Michigan pilot will test whether the new tools can identify those lower-risk fields and tailor more precise insurance and lending products, including broader coverage and potentially lower-cost options for growers.

Where the safety net still applies

For most catastrophic losses, farmers will still rely on USDA coverage. The op-ed focuses on "shallow losses" — the smaller but frequent 0% to 30% losses measured against historic yields. Field-specific, AI-informed products can offer broader, more tailored protection in that band. Some growers may pay more or differently because it better protects their operations and gives them greater freedom to manage risk beyond existing products. Taxpayers could save as growers adopt practices that lower risk and draw on fewer federal subsidies and indemnity payments.

The farm bill window

The Senate Agriculture Committee passed its draft farm bill out of committee last week. Trade disputes and international conflict have compressed prices and lifted input costs, intensifying demand for safety net programs. The emerging field-level products are too new to influence current deliberations, but Bonnie and Sherrill argue the implications should not be lost on policymakers. They point to the broader payoff: "more precise and potentially broader protection for farmers and ranchers, savings for taxpayers, and better-managed land."

What comes next

The partners plan to expand the model across the Corn Belt and, eventually, to other crops. The technology now exists to price a single field differently, the authors wrote, and a pilot launching in four Michigan counties will show whether the rest of the country is ready to follow.