What changed
Two concrete rules just moved eligibility boundaries: USDA made previously-ineligible disaster losses (unborn livestock, bird depredation, expanded drought) claimable at better rates (FACT, cited), and DHS granted SLTT police/corrections agencies brand-new counter-UAS authority bundled with training/certification/reporting obligations (FACT, cited). Separately, Cloudflare shipped x402 per-request machine payments (FACT, cited).
Why now
Rules routinely mint whole new eligible or obligated populations overnight, and most affected parties don't read the Federal Register. Free ingest APIs (Federal Register, Grants.gov) plus cheap LLM classification make an eligibility-expansion diff engine buildable by one person today; the value is being FIRST to the freshly-eligible party's advisor.
Converging signals
A money-expanding rule (USDA 5585), an obligation-creating rule (counter-UAS 5638), and a per-request monetization rail (x402/16) meet at one point: a metered feed of 'the criteria just moved, here's who it affects.' The x402 rail is the weakest leg — it is a nice-to-have, not the business.
Customer pain
HYPOTHESIS (not proven by supplied evidence — demand_evidence is empty): grant/claims consultants and ag lenders need fresh leads of newly-qualifying clients and today rely on stale, generic databases and manual Federal Register reading. Existing spend of $124-250/mo on grant databases is asserted in the leap, not in the source signals — treat as unverified.
Who pays
Buyer ≠ beneficiary. Beneficiaries are producers/agencies who gain money or duties. The BUYER is the professional who monetizes them: grant/farm-loan consultants, ag lenders, and compliance vendors selling certification/reporting services to counter-UAS agencies. They pay for fresh, pre-qualified leads and obligation alerts.
Solved today
Free Federal Register email/RSS alerts and keyword search; Regulations.gov; paid reg-trackers (FiscalNote/Regology, Bloomberg Government, POLITICO Pro); grant databases (Instrumentl, GrantWatch). None specifically diff for 'eligibility EXPANDED / new obligated class' and match to named parties.
Why current solutions are bad
Generic keyword alerts flood the user with irrelevant rules and never answer 'who NEWLY qualifies.' Grant databases list programs, not rule-level eligibility changes. The gap is the eligibility-DELTA classification plus party-matching — but it is a thin, copyable layer.
Proposed product
A daily eligibility-delta digest + queryable API: ingest final/proposed RULES only (filter out Notices), LLM-classify each as newly-claimable-money vs new-obligation, extract the affected population + dollar/loss criteria, join to public entity lists (USDA producer/county data, SLTT agency rosters), and surface 'X party class just became eligible/obligated.' Sell as a subscription to a single vertical first.
MVP version
Pick ONE vertical — USDA/FSA disaster & loan eligibility — and one buyer, ag/farm grant consultants and ag lenders. Cron the Federal Register API (type=RULE/PRORULE) + Grants.gov, run a classifier prompt tagging eligibility-expansion and extracting criteria, produce a curated daily email digest with source links and a plain-English 'who newly qualifies' summary. Human-in-the-loop review before sending to protect accuracy. Skip x402/MCP for MVP; bill by card.
30-day build
Build ingest + classifier for the ag/FSA vertical; hand-curate a 4-week backlog of real eligibility-expanding rules; run the KILL TEST — take 20 such rules and contact affected parties/their consultants to measure how many already knew. If >70% already knew, pivot or kill.
60-day build
If freshness survives the kill test, ship the daily digest to 10-20 design-partner consultants/lenders free; instrument open/click and 'was this a lead you didn't have' feedback; harden classifier precision (false 'you're eligible' alerts are fatal to trust).
90-day revenue plan
Convert design partners to paid at $99-299/mo; add a second vertical (broad grant-eligibility or the counter-UAS compliance-obligation feed sold to certification/reporting vendors). Target first paid revenue by ~day 75-90.
Distribution path
Direct outreach to ag lenders and farm/grant consultants; presence in grant-writer and Farm Service Agency-adjacent communities/newsletters; a public 'this week's eligibility changes' teaser digest as a top-of-funnel lead magnet.
Pricing hypothesis
$99-299/mo per seat by vertical; optional metered API/MCP endpoint later for programmatic buyers (x402 only if a real machine-buyer appears). Contingency/referral fees from closed claimants are possible but add licensure/ethics complexity — avoid at first.
Technical difficulty
Low-to-moderate: free APIs, standard LLM classification, entity-list joins. The hard part is classification PRECISION and party-matching accuracy, not infrastructure.
Legal / regulatory risk
Low if positioned as information/lead product. Do NOT give legal/eligibility determinations as advice; label outputs as leads to verify. Contingency-fee or claim-filing models would trigger finder-fee caps and possible licensure — out of MVP scope.
Platform dependency
None on a government platform (no deplatform risk). Dependent on Federal Register/Grants.gov API availability (stable, free). x402/Cloudflare dependency avoided by deferring metering.
Founder fit
Strong. This is the founder's primary thesis — public money flows + regulation changes + a paperwork/awareness gap — and reuses his government-portal, public-records, and complaint/data-mining strengths. It is a data/alert product, not enterprise software or a network-effect play.
Breakout potential
Moderate-to-high IF the freshness edge is real: one working vertical replicates to dozens (every agency's money + every new obligated class + all 50 states' pass-through programs). But the core diff is copyable by incumbents, so the moat is speed, vertical depth, and party-matching data — not the concept.
Final recommendation
CONDITIONAL BUILD. The shape fits the founder perfectly, but the whole thesis rests on one unproven assumption (under-noticed eligibility changes). Do NOT build the broad 'Federal Register mempool' — narrow to USDA/FSA ag eligibility sold to ag lenders/farm consultants, and gate the build on passing the kill test first. Drop x402 from MVP.
Next action
Run the kill test now: pull 20 eligibility-EXPANDING rules from the last quarter, contact the affected parties (or their consultants), and measure what fraction already knew before building anything.