What changed
FACT (from source text): DHS/FEMA obligated $2,350,410,840.81 to the North Carolina Department of Public Safety to 'PROVIDE REIMBURSEMENT TO STATE, LOCAL, TRIBAL, AND TERRITORIAL GOVERNMENT ENTITIES AND CERTAIN PRIVATE NON-PROFIT ORGANIZATIONS FOR EMERGENCY PROTECTIVE MEASURES TAKEN DURING THE PANDEMIC,' explicitly naming emergency medical care, medical sheltering, and vaccine administration/distribution. INFERENCE: the money is a pass-through — NC DPS is the recipient of record, but the actual filers are the subrecipient local governments, tribes and private non-profits who must document costs to draw the money down and to survive later audit.
Why now
HYPOTHESIS, and the weakest link in this idea: the award text describes PANDEMIC emergency protective measures. Federal COVID-19 disaster declarations were issued in 2020; by mid-2026 the overwhelming majority of Category B COVID projects are obligated, and many are in closeout or already de-obligation/appeal territory. Nothing in the provided source states an open application deadline — the input itself records 'DEADLINE: none stated in text.' So the 'why now' is NOT 'a filing window just opened.' The honest 'why now' is the audit tail: obligated dollars invite OIG and state single-audit scrutiny, and unsubstantiated costs get clawed back. That is a real but narrower, slower, and more adversarial market than a fresh filing mandate.
Converging signals
Three provided signals genuinely converge: (1) a very large obligated FEMA PA award with a named eligible-claimant class (FACT, usaspending 4827DRNCP); (2) other multi-billion-dollar DHS PA awards to other states for disaster damage — Puerto Rico $35.3B, New York $17.4B, Virgin Islands $22.0B (FACT, usaspending) — proving this claimant/paperwork structure repeats across states and disasters, which is what makes the product replicable; (3) the FEMA Grants Portal as the fixed submission surface. The Medicaid entitlement awards in demand_evidence (HHS T19, NC/NY/CA/FL/etc.) are semantic false positives — they are formula entitlements to state agencies, not a filer class doing cost substantiation. I am explicitly declining to score demand from them.
Customer pain
HYPOTHESIS (not evidenced in the provided input): a PNP hospital or a county finance office must tie every claimed dollar — overtime differentials, contract nurse invoices, PPE POs, vaccine-clinic staffing hours — to a specific eligible emergency protective measure under 44 CFR 206.225, in a narrative FEMA reviewers accept. Their source data lives in payroll and ERP systems that were never built to tag costs by disaster or by eligibility category. Reconstruction is manual, and errors surface years later as de-obligations. No complaint threads, job postings, or PRA respondent counts in the input corroborate this pain. I am labeling it inference, not fact.
Who pays
Best case: the PNP hospital's grants/reimbursement manager or a county emergency manager / finance director, with a $3k-$8k per-project budget line that is itself often reimbursable as administrative cost. Realistic case: the buyer is not them — it is the FEMA PA consultancy that already holds the engagement and would rather build or license than refer. Selling to the consultant as a tooling vendor is a smaller but far more reachable business than selling to hospitals.
Solved today
Contingency and hourly consultants: Hagerty Consulting, Witt O'Brien's, Tidal Basin, IEM, and the large accounting firms run FEMA PA cost recovery as a practice line, typically billing hourly or a percentage of recovered/sustained funds. States also fund subrecipient technical assistance. FEMA supplies the Grants Portal itself for free, so the submission surface is not the bottleneck — the substantiation and narrative work is.
Why current solutions are bad
Consultant fees consume a meaningful slice of the award, and the consultant's spreadsheet knowledge walks out the door at engagement end, leaving nothing the claimant can re-run when the auditor arrives three years later. That is a genuine gap. But note the counter: consultants are not merely doing data entry — they are absorbing professional judgment and reputational risk on eligibility determinations. Software that outputs a determination without standing behind it does not replace them.
Proposed product
A cost-substantiation workspace, not a submission bot. Ingest payroll registers, PO/AP exports, and executed contracts as CSV/XLSX; apply a rules engine keyed to FEMA PA eligibility categories with per-line provenance; flag ineligible or duplicative costs before they are claimed; and export a versioned audit binder (cost schedules, source-document index, eligibility narrative per cost line) plus a Grants Portal-ready project worksheet package. The durable asset is the binder that survives personnel turnover and reproduces the claim on demand at audit.
MVP version
Single disaster, single claimant type. Payroll + AP CSV ingest with a mapping UI, an eligibility rules engine encoding 44 CFR 206.225 and the current FEMA PA Program and Policy Guide for Category B, per-line eligibility flags with citations to the policy paragraph, and a PDF/XLSX audit binder generator. No portal API integration in v1 — FEMA Grants Portal has no public partner API I can assert exists from the provided sources, and building against an undocumented authenticated portal for a claimant class filing a handful of projects is not where the value is. Manual upload of the generated package is acceptable to the buyer.
30-day build
Do not write the rules engine yet. Interview 15 people: five county emergency managers, five PNP grants managers (start with volunteer fire/EMS PNPs, where the founder's fire-service background gives him a real hearing), and five FEMA PA consultants. The single question to answer: are there open Category B projects still being substantiated, or is this entirely a closeout/audit-defense market? Pull the FEMA PA subrecipient obligation data for DR-4827 and adjacent disasters to count claimants and check project status. If the answer is 'closeout only,' kill the COVID framing on day 30 and re-aim at the current disaster year.
60-day build
Build the Category B rules engine against one real (anonymized) historical claim supplied by a design partner, and validate that the tool's eligibility flags match what FEMA actually obligated on that project. This backtest is the entire technical risk: if the engine cannot reproduce a known-good determination, there is no product. In parallel, negotiate a paid pilot with one consultancy as a tooling license.
90-day revenue plan
First revenue from two to four paid pilots at $3,500-$6,000 per project (flat, not contingent), or one consultancy tooling license at $1,500-$2,500/month. This is a 120-180 day path realistically, not 90. Reject the '% of reimbursement sustained on audit' model outright: audit sustainment resolves on a three-to-seven-year horizon, would make the founder an unsecured creditor of his own customers, and edges toward contingency-fee grant consulting, which several states regulate.
Distribution path
State emergency management association conferences and the state PA officer network; direct outreach to PNP fire/EMS agencies where the founder's fire-service credibility converts; and channel sales to the mid-tier PA consultancies. Content wedge: publish a free 'Category B de-obligation post-mortem' analyzing published OIG audit reports on FEMA PA recipients, showing exactly which cost lines got clawed back and why. That artifact demonstrates value the way this founder actually sells.
Pricing hypothesis
$3,500-$6,000 per project for direct claimants; $1,500-$2,500/month per seat for consultancies; $10k-$25k for a state-level subrecipient portal deployment (this last is enterprise procurement — treat it as an outcome, not a channel).
Technical difficulty
The software is easy; the content is hard. CSV ingest, a rules engine and a PDF generator are a few weeks of AI-assisted work. Encoding FEMA PA eligibility so that its determinations match what FEMA and OIG actually sustain — that is expert domain knowledge the founder does not currently have, and it cannot be bootstrapped from the PAPPG alone. The ERP export mapping (Workday, Lawson, Tyler, Epic) is a real integration tax: every claimant's payroll export is shaped differently.
Legal / regulatory risk
Moderate and understated by the input's framing. Software that outputs an eligibility determination on a federal cost claim is producing an assertion the claimant certifies to the government. If the engine flags an ineligible cost as eligible and it is claimed, the claimant is exposed and will look upstream. Requires disclaimers, a decision log, and E&O coverage. The proposed '% of reimbursement' fee compounds this by aligning the vendor with over-claiming. Not disqualifying, but this is not the risk profile of the founder's ELDT upload tool.
Platform dependency
Low and correctly so. The submission target is a government system, and no platform owner can deplatform this. Not flagging platform_policy_risk. The real dependency is on FEMA policy stability — the PAPPG is reissued periodically and the rules engine must track it, which is ongoing maintenance, not a kill.
Founder fit
Genuinely strong on the shape and weak on this specific instance. The public-money/forced-filer pattern is exactly the founder's thesis, and his fire-service background gives him authentic access to PNP fire/EMS claimants — that is a real, unusual distribution edge. But this is not the ELDT shape. ELDT was a per-transaction upload against a fixed schema with a recurring, forward-looking filer population. Category B COVID substantiation is a one-time, backward-looking, judgment-heavy reconstruction for a claimant class that has largely already filed. The transferable asset here is the audience, not the analogy.
Breakout potential
Bounded as scoped, real if re-aimed. Every declared disaster regenerates the same Category B/E paperwork for a fresh set of claimants — the Puerto Rico, New York and Virgin Islands PA awards in the evidence set (FACT, usaspending) show the structure repeating at scale across states and years. A tool aimed at the CURRENT disaster year, sold once and re-run every declaration, compounds. A tool aimed at pandemic closeout does not.
Final recommendation
CONDITIONAL PIVOT — do not build the idea as stated; build its forward-looking sibling. The public-money shape is right and the founder's PNP fire/EMS access is a genuine edge, but the specific mandate is expired, the monetization is unsound, and the incumbents sell risk absorption rather than labor. Re-aim at Category B/E cost substantiation for the CURRENT disaster year, sold flat-fee to small local governments and PNP fire/EMS agencies — the claimants too small for a consultancy to bother with, where the founder's credibility lands and where there is no incumbent. Gate the build on a single 30-day finding: are there open, unfiled PA projects in the current declaration year with claimants who have no consultant. If yes, this is an A-grade opportunity in a different disaster. If no, kill it. The evidence provided does not answer that question, and no amount of reasoning over these nineteen Medicaid awards will.
Next action
Pull the FEMA Public Assistance subrecipient dataset for the three most recent major disaster declarations, count claimants with obligated-but-not-closed Category B projects and under $500k in total obligation (the consultant-free tier), and cold-call ten of them from a fire-service introduction. Decide by day 30 on that count alone.