Portfolio-matched public-record enrichment
A scoped, paid engagement to evaluate — on your own book — whether public-record events (SBA credit outcomes, FMCSA operating-authority status, OSHA workplace-safety enforcement, EPA environmental enforcement, openFDA product recalls, USPTO IP collateral, SBA PPP loans, Form 5500 identity, federal award dependency and SEC presence) add decision-relevant signal to your counterparty assessment. No platform to adopt, nothing to integrate to decide.
1What you provide
A list of the US counterparties you want assessed. Dirty legal names are fine — resolution is our job, not yours.
| Field | Required | Note |
|---|---|---|
| your_id | yes | your opaque key, echoed back untouched |
| legal_name | yes | as you hold it — abbreviations / suffixes handled |
| ein / cik / ticker / duns / lei | optional | any identifier sharpens matching to near-certain (an EIN is the strongest anchor) |
| state / country / domain | optional | corroborating signals when present |
| anonymized outcomes | optional | downgrade / claim / default / limit change — lets us compare our events to your own outcomes |
Delivered as CSV or JSONL over an encrypted channel. Typical pilot size: 500–5,000 names.
2What you receive
- —The matched file — one row per counterparty: the resolved entity, a status (auto / review / unmatched) and an explainable confidence, plus a name-match vs entity-confidence read (an exact name is not, on its own, a confirmed identity).
- —Per-register findings, attached to each matched name where the register applies: SBA loan-level credit outcomes (charge-offs / pay-offs, dated); FMCSA operating-authority status and insurance-lapse / suspension sequences; OSHA high-severity workplace-safety enforcement (willful / repeat violations); EPA high-severity environmental enforcement (formal CAA/CWA/RCRA actions / significant non-compliance); FDA high-severity product recalls (Class I or still ongoing, across drug / food / device); USPTO patents pledged as loan collateral (the lender, the recordation, the dates); Form 5500 EIN-grounded identity + name variants; Federal awards contract/grant dependency (USAspending) with SAM.gov registration & exclusion status; SEC presence and 8-K / CMBS signals.
- —A consolidated dossier per counterparty — every register on one point-in-time page, with a cross-register timeline and the source link behind each finding.
- —An enrichment report — coverage on your book by register and recency, and the highest-signal cases. See a sample enrichment report →
- —An evaluation pack you verify independently: a signed sample, the data dictionary, and scripts that re-check every signature and re-join the data on a clean machine.
A distribution SME
a dirty name in your file, resolved and enriched
- Identity — Form 5500EIN-grounded · NAICS 42 · ~40 active participants
- SBA credit2 loans found · 1 charged off (2023) · 2 of 2 outcome-observable
- FMCSAno confident match
- OSHA3 inspections · no willful/repeat violations
- EPA2 regulated facilities · no formal enforcement action
- USPTO IPno patent security interests
- Federal awardsno federal contracts on record
- SECno confident match
The signal is the SBA charge-off — a late, loan-level outcome, not a company default. Entity confidence: medium (name + state).
A regional motor carrier
the same resolution, a different register lights up
- Identity — Form 5500EIN-grounded
- FMCSAoperating authority SUSPENDED after an insurance lapse (2024)
- OSHA1 willful safety violation cited (2023)
- EPA1 formal enforcement action (Clean Water Act, 2022)
- SBA credit1 loan · paid in full
- USPTO IP / Federal / SECno confident match
The signal is the FMCSA authority suspension — it bars the interstate for-hire operations covered by that authority while the suspension remains in force.
THORNYCROFT GILES AND COMPANY
8 patent-collateral events between 1997 and 1999 — the lender, the recordation and the patents on each. View the complete timeline →
A factual report of public USPTO records.
3How matching & signal are validated
- —Explainable, deterministic matching — external-id, exact name, or fuzzy; every resolution carries its confidence and method. No black box, no ML guess.
- —Precision over coverage — audited on a stratified sample (matching, lender normalization, and event-delta classification each scored against a manually verified gold set). We report measured precision, not assertions.
- —Abstention by design — an entity we cannot resolve with evidence stays unresolved under its exact raw name. We never force a merge to inflate a number.
- —Point-in-time, per source — filing / recordation events (USPTO, SEC) dated as of their public record date; historical SBA outcomes are retrospective labels, with vintage capture from 2026 onward. Failed / delisted entities are retained (no survivorship bias); each event replays to its source document.
4Data handling & security
- —Data minimization. We need names + identifiers, nothing about your exposures, limits or pricing.
- —Encrypted in transit and at rest; access limited to the operator running the pilot; no third-party processors beyond the hosting infrastructure.
- —Deleted at the end of the pilot (or on request), unless you choose to continue. Your list is never redistributed, resold, or used to train anything.
- —A data-processing agreement is available; processing location and retention terms are fixed in writing before any file is sent.
5What success looks like
The pilot is a measurement, not a demo. We agree the bar up front; these are the axes we score.
| Measure | What it answers |
|---|---|
| Coverage on your book | share of your counterparties with ≥1 finding in any register, on recent years |
| Novelty | share of findings you did not already hold in your credit file |
| Audited precision | correctness of the match and the per-register findings on a stratified reviewed sample |
| Timeliness | for each finding, the gap between the public record date and the date it reached you — how much of your review window we give back |
| Decision relevance | on named cases, would the signal have changed or informed a call? |
If the honest answer is « not enough », that is a valid, cheap result — and you have lost nothing but a scoped engagement.
6Timeline
| Phase | Weeks | Output |
|---|---|---|
| Scoping & DPA | 0–1 | fields, success bar, data terms agreed |
| Match & enrich | 1–3 | matched file + enrichment report |
| Joint review | 4–6 | audited precision, novelty, named cases |
| Decision | 7–9 | go / no-go on a recurring feed |
7Beyond the pilot
If the signal proves its worth, the pilot converts to a recurring enrichment feed — your book re-scored as new records post across every register. Candidate extensions, only if you ask: further public registers, trademark security interests, and delivery into your existing workflow.
Scope a pilot — contact@redfleet.fr
Factual summaries of public US records (SBA, FMCSA, DOL OSHA, EPA ECHO, openFDA, USPTO, DOL Form 5500, USAspending, SAM.gov, SEC). A filing, charge-off, award, citation, recall or regulatory status is not an allegation of wrongdoing or a prediction of company default. Not affiliated with any government agency. Not investment or legal advice.