How we keep DueVestor reliable

Sanctions coverage — every report, every time

Every DueVestor report screens the full sanctions catalog — 22+ official government and multilateral lists. The four anchor lists below are the always-on core; their order is reshuffled for your region, but the catalog is identical for every customer — presentation can reorder, the catalog cannot lose a connector.

• ofacOFAC SDN — U.S. Treasury Specially Designated Nationals.
• unUN Consolidated — UN Security Council sanctions.
• euEU FSF — EU Consolidated Financial Sanctions list.
• ukUK OFSI — UK Office of Financial Sanctions Implementation.

Verification methodology

Four rules govern every finding in every DueVestor report. They are the difference between an aggregator’s mirror and a dossier you can defend before a board, a bank, or a regulator.

Multi-source screening by default

Every subject is screened in parallel against 144+ official data sources — sanctions, watchlists, corporate registries and court records drawn directly from government and multilateral publishers, never a single aggregator's copy — with checks routed across 38 jurisdictions and corporate-registry depth in 35+ markets.

A citation on every finding

Each finding carries its source, the retrieval timestamp, and a confidence grade. Reports are deterministic and replayable — your auditors get a chain of custody, not a black box.

Confidence tiers are earned, not asserted

Findings are graded LOW / MEDIUM / HIGH. HIGH is reserved for findings independently corroborated by a second, unrelated source. A single-source finding stays MEDIUM no matter how authoritative the source — and a sanctions hit is never auto-dismissed as a false positive.

A transparent coverage statement in every report

Every dossier closes with a sources-screened table: what was checked, what answered, and what was skipped or unreachable — with cache age where a fallback served. If coverage was partial, the report says so. We never imply coverage we did not have.

Cross-verification: while our second, commercially independent screening database is being queried, every sanctions and PEP finding is checked against it as well as against our direct official-source pipeline, and a finding is graded HIGH confidence only when both agree. That second source runs on a metered plan with a monthly ceiling; in a run where it is not queried, the report names it as not screened and no finding in that run is presented as cross-verified.

What we monitor

Every external data source DueVestor depends on — OFAC, the UN consolidated list, the EU FSF, UK OFSI, the World Bank debarred-firms list, SEC + DOJ FCPA enforcement, Wikidata PEP, GDELT adverse media, and court-record sources (CourtListener plus jurisdiction-specific sources routed per subject domicile) — is probed every 60 minutes with a known-safe query that should always produce hits. Every result is recorded in our audit trail, alongside every report-driven fetch. The dashboard you see at /status reads from the same audit channel — there is no separate marketing-only data path.

How we self-heal

• • Transient errors retry with exponential backoff (5s → 30s → 2 min, up to 3 retries).
• • After five consecutive failures, a circuit breaker opens for that source — we stop hammering an upstream that is clearly broken. A 30-minute recovery window auto-re-tests; success closes the circuit, failure reopens it.
• • For sanctions lists, we keep the last-good parsed record list and serve from it while the circuit is open — your report still gets evidence-grade hits, marked with the cache age so you know exactly how fresh the data is.
• • Every finding emitted by a connector is checked against peers. When two independent sources confirm the same entity, the finding is tagged HIGH confidence; single-source findings are MEDIUM.

Transparency

We publish a real-time status page + an incident log. Any outage of 30+ minutes auto-creates a public incident; resolution + post-mortem notes land on the same page. You can embed our status badge on your own site if you want.

The same rule governs our marketing: every coverage number we publish is generated from the live connector registry and rounded down — claims update as coverage grows, never ahead of it. And every report states exactly which sources were checked, with a citation for every finding.

Our target

99.5% monthly uptime on report generation. Sources we do not control may degrade — when that happens, we will surface it in the report itself (with the source state and cache age) and tell you on the status page. We will never silently serve stale or incomplete data without flagging it.