What changed
FACT: A USDA final rule effective 2026-07-09 (Federal Register 2026-13878) expanded Supplemental Disaster Assistance eligibility — new claimable loss types (unborn livestock, bird depredation) and relaxed drought conditions, plus higher marketing-assistance-loan rates. INFERENCE: a concrete, dated population of producers just became owed money they previously could not claim.
Why now
FACT: the rule is fresh (published 2026-07-09), so the awareness window is open right now. FACT (capability): schema-defined structured extraction from Federal Register/grants.gov is now a single API call (Context.dev, YC S26), collapsing the build cost of a rule-diffing monitor. HYPOTHESIS: producers learn of expansions slowly via FSA/extension offices, creating a time-arbitrage window before consultants catch up — this is the core unproven assumption.
Converging signals
Three signals meet: (1) a rule that quietly lowers an eligibility bar and creates a forced-eligible class, (2) the same 'you-qualify-now-but-don't-know' gap generalizes to FEMA PA subrecipient reimbursement and other programs, (3) one-API-call structured extraction making the diff engine solo-buildable. The convergence is the DELTA between rule versions, not another static grant database.
Customer pain
FACT (from signal 2388 referenced in input): existing $124–250/mo static grant databases are 'hated' — they list everything and surface nothing timely or actionable. HYPOTHESIS: lenders/co-ops lose repayment capacity and advisors lose success fees when borrowers/clients miss claim windows they never knew opened. No demand_evidence array was provided, so this pain is asserted, not evidenced.
Who pays
Primary buyer (the BUYER, not the beneficiary): ag lenders, farm credit co-ops, and disaster-recovery/grant advisors who profit when their borrowers/clients capture payments — subscription. Secondary: success-fee referral on assembled claims (subject to caps/rules). The beneficiary (the producer) is NOT the initial buyer. INFERENCE.
Solved today
Static grant/database subscriptions, FSA county-office outreach, agricultural extension bulletins, and percentage-of-award consultants who assemble claims. HYPOTHESIS: none of these diff rules the day they publish and push a geo-targeted 'you are newly eligible' alert.
Why current solutions are bad
Databases are comprehensive but not timely or personalized — they don't tell a specific producer in a specific county that a bar dropped today. Consultants are expensive and only engage after the producer already suspects eligibility. The awareness gap between publication and action is the unserved seam.
Proposed product
A rule-change 'eligibility-expansion' tripwire: ingest Federal Register RULE/PRORULE + grants.gov via structured extraction, diff each rule against its prior version for lowered thresholds / added loss categories / rate increases, tag by geography and NAICS, and fire email/SMS 'you can now claim X' alerts plus a single-program intake wizard (start with USDA disaster ag). White-label the alert feed to lenders/co-ops.
MVP version
Manual-assisted diff of the 2026-07-09 USDA rule against the prior program rule; hand-build the newly-eligible profile (loss types, geographies, NAICS); an intake wizard for that ONE program; and an alert template. Then wire Context.dev extraction + a nightly Federal Register RULE-only poll to automate the diff. Land 1-2 lenders/co-ops as design partners before building breadth.
30-day build
Run the KILL TEST first: take the actual 2026-07-09 USDA expansion and call ~10 affected producers in one county — if most already know they're newly eligible, the awareness gap doesn't exist. In parallel, interview 5 ag lenders/co-ops and 3 disaster advisors on whether they'd pay to alert borrowers. Build the single-program intake + diff prototype only if the awareness gap and buyer willingness both hold.
60-day build
Automate Federal Register (filtered type=RULE/PRORULE) + grants.gov ingestion and the version-diff engine; add geo/NAICS tagging; ship white-label alert feed to 1-2 design-partner lenders/co-ops; instrument open/claim-start rates to prove the arbitrage window empirically.
90-day revenue plan
Convert design partners to paid subscriptions (per-seat or per-institution); add a second program (FEMA PA subrecipient or a state pass-through disaster program) to prove replication; explore compliant success-fee referral on assembled claims where finder-fee rules permit.
Distribution path
Direct outreach to ag lenders, Farm Credit associations, and co-op ag advisors; farm-lending and disaster-recovery associations; demonstrated-value demo using the live 2026-07-09 USDA expansion as the proof case. Not consumer producer marketing initially — sell through the adjacent professional who already has the borrower/client list.
Pricing hypothesis
HYPOTHESIS: $150–500/mo per lender/co-op branch for the alert feed + intake, undercutting the hated $124–250/mo static databases on relevance while priced per-institution not per-producer. Optional compliant success-fee on assembled claims where lawful. Validate against actual willingness-to-pay in the 30d interviews.
Technical difficulty
Low-to-moderate for a solo AI-assisted dev: structured extraction is one API call; the real work is the version-diff logic (detecting a lowered threshold / added loss category reliably, not just text changes) and accurate geo/NAICS matching of the eligible population. Alert/intake plumbing is standard.
Legal / regulatory risk
Moderate. Success-fee/finder-fee models can be capped or restricted for benefit claims — verify per program/state before charging producers. Do NOT give eligibility 'advice' that crosses into unauthorized practice; frame as informational alerts + intake, not legal determination. No licensure needed to run an alert/monitor SaaS sold to lenders. Avoid predatory monetization of producers.
Platform dependency
Low. Sources are public government systems (Federal Register, grants.gov) with no platform owner who can deplatform. Dependency on the Context.dev extraction API is a swappable vendor risk, not a policy risk. SMS via a standard provider.
Founder fit
Strong. This is the founder's primary thesis shape — public money + a rule change creating a forced/eligible class + a paperwork/awareness gap + a government-data monitor with per-seat monetization. Matches his public-records, compliance-monitor, and government-portal-integration edge (FMCSA ELDT). The wrinkle: the buyer is the adjacent professional, and the core assumption (awareness gap) is unproven and must be kill-tested.
Breakout potential
Real if the awareness-window thesis holds: one working program replicates across USDA programs, FEMA PA, and 50 states' pass-through disaster/grant programs — the same diff engine, new eligibility profiles. The moat is the maintained library of rule-version diffs and eligibility profiles, which compounds over time.
Final recommendation
CONDITIONAL PURSUE — validate before building. The shape is an excellent founder fit and the 2026-07-09 USDA rule is a concrete, dated proof case, but demand_evidence is empty and the whole idea lives or dies on an unproven awareness-arbitrage window and an unvalidated buyer. Spend the first 30 days on the kill test and buyer interviews, not code. Build only if both the awareness gap and lender/co-op willingness-to-pay survive.
Next action
Execute the KILL TEST: pull the actual newly-eligible producer profile from the 2026-07-09 USDA rule, pick one high-impact county, and call ~10 affected producers to measure whether they already know they're newly eligible — simultaneously interview 5 ag lenders/co-ops on paying to alert borrowers.