What changed
On 2026-07-09 USDA published a final rule (Supplemental Disaster Assistance Programs, Marketing Assistance Loans, and Sugar Provisions) that expands disaster assistance to previously-ineligible loss types — cited examples include unborn livestock and bird depredation — and lowers drought eligibility thresholds, making some past losses retroactively claimable at higher rates. (FACT: Federal Register 2026-13878.)
Why now
The eligibility expansion is fresh and most producers are unaware they now qualify (HYPOTHESIS — the input asserts low awareness but provides no direct evidence). Simultaneously, schema-defined structured extraction (context.dev) and cheaper frontier reasoning (the GPT-5.6 cost-performance claim) make it economically viable for a solo dev to mine messy farm records and cross-reference public drought/satellite data — the exact assembly work that was previously too labor-intensive to productize. (HYPOTHESIS on the capability-cost link.)
Converging signals
Three signals meet: (1) a new federal rule creating retroactively claimable money (money), (2) one-call structured extraction from arbitrary records/sites (dev), (3) frontier-model reasoning cheap enough to reconstruct and substantiate events from proxy data (ai). The scarce good is not the claim form — it is defensible evidence of a loss that was invisible until the rule changed.
Customer pain
Latent, not felt: producers don't know they're newly eligible, lack contemporaneous documentation for events (a stillborn calf, depredated poultry, a drought-thinned herd) they never bothered to record because it wasn't claimable, and face a filing deadline. There is NO demand_evidence in the input (no complaints, no job ads, no forced-buyer mandate) — this is a discretionary claim, not a compelled filing, so felt urgency is low until awareness is created.
Who pays
Two candidate buyers: (a) farmers/ranchers on a success fee, and (b) ag lenders / farm-credit associations / co-ops who benefit when their borrowers recover federal payments (improves loan performance). The lender/co-op channel is the stronger buyer — reachable, repeat, and it solves the farmer-acquisition problem for you.
Solved today
FSA county offices are staffed and mandated to help producers file disaster claims for free; crop-insurance agents and some ag consultants assist with paperwork. Livestock Forage Program payments are already auto-triggered by Drought Monitor designations, so part of the 'reconstruction' thesis is redundant for drought-forage losses.
Why current solutions are bad
County FSA help assumes the producer knows to show up and has documentation; it does not proactively surface newly-eligible retroactive events buried in a farm's own breeding/vet records, and it does not assemble proxy substantiation from public drought/satellite data. That gap — retroactive discovery + evidence manufacture — is the wedge.
Proposed product
An ingest-and-assemble service: pull a producer's herd/breeding/vet records and photos, detect events now matching the new eligibility categories, overlay dated USDM drought designations and satellite indicators, and output an FSA-format substantiation packet with a claimable-value estimate and per-program deadline tracker. Delivered white-label to lenders/co-ops rather than sold one farmer at a time.
MVP version
A single-program, single-region pilot: ingest one cooperating producer's records for the highest-value newly-eligible category (e.g. livestock indemnity for unborn/depredation losses), auto-match USDM dates, and produce one FSA-ready packet — then run the KILL TEST by filing it.
30-day build
Read the final rule and the FSA handbooks for the affected programs (LIP/ELAP/NAP/LFP) to confirm EXACTLY what documentation FSA accepts. Recruit one cooperating producer and one friendly county-office or ag-consultant contact. Build the extraction + USDM-overlay prototype against that producer's real records.
60-day build
File the reconstructed packet through the cooperating producer. This is the gating event — do not scale before FSA accepts proxy substantiation. In parallel, pitch 2-3 ag lenders/co-ops on a white-label pilot contingent on the acceptance result.
90-day revenue plan
If the packet is accepted: sign one lender/co-op to a paid pilot (per-packet or per-recovered-dollar fee) and process a first batch of their borrowers against the filing deadline. If rejected: pivot to a pure awareness/eligibility-screening data product (no substantiation claim) or shelve.
Distribution path
B2B2C through ag lenders, farm-credit associations, cattlemen's/commodity groups, and co-op extension networks — they have the borrower/member lists and a financial interest in recovery. Direct-to-farmer is a poor channel (high CAC, skepticism of middlemen).
Pricing hypothesis
Per-packet fee ($500-$2,500 depending on program) or a capped success fee, sold to the lender/co-o co-op as a white-label per-member service. Avoid uncapped percentage-of-award framing to sidestep finder-fee/representation concerns; price the documentation work, not the money.
Technical difficulty
Moderate. Extraction and USDM overlay are tractable with the cited tooling. The hard part is FSA-format correctness and defensibility of proxy evidence — accuracy, not engineering, is the risk.
Legal / regulatory risk
Moderate. Helping a producer document their own losses is generally not licensed practice, but representing claimants before USDA for a contingent fee, and any percentage-based success fee, can implicate representation and finder-fee rules — get the fee structure reviewed and keep the producer as the filer of record.
Platform dependency
None on a private platform — output goes to a government system (FSA), which cannot deplatform you. Dependency is on FSA's documentation acceptance policy, not a marketplace owner.
Founder fit
Good but not perfect. It sits in his claimable-money + public-records + data-product wheelhouse and leverages ops/records thinking, but it is a success-fee / managed-service shape, NOT the proven per-filing portal-submission pattern behind his FMCSA ELDT app. Founder fit is real but lower than a pure forced-filer mandate.
Breakout potential
If the kill test passes, it replicates across ~50 states, multiple disaster programs, and every future disaster designation — a recurring, expandable engine. If it fails, it collapses to a low-value awareness tool.
Final recommendation
CONDITIONAL BUILD — worth a cheap, fast validation but do not scale on faith. Spend <30 days and near-zero capital to run the kill test: file exactly one reconstructed packet. Gate everything on FSA acceptance and on signing one lender/co-op channel partner. High upside if both clear; walk away cleanly if either fails.
Next action
Read the FSA program handbooks for the affected disaster programs to confirm accepted documentation, then recruit one cooperating producer and file a single reconstructed packet as the kill test.