None of them exotic. All measured in cash. All tractable with the right evidence base.
Overbroad classifications. Costs treated as dutiable that never were. Exemption headings nobody claimed. Refund windows that closed quietly. Every one of these is a line in your entry summaries — visible, provable, recoverable, if somebody reads them all. Nobody reads them all by hand.
Measures take effect overnight. The decisions they force — hold the container, re-route the order, re-price the SKU, file inside the in-transit window — expire in days. An answer that arrives in three weeks is not an answer. It is a post-mortem.
Trade enforcement has moved up the priority list, and reasonable care is judged on documentation, not intention. A classification with no written reasoning, no cited rulings and no timestamp is an exposure waiting for a focused assessment.
Ten-digit lines with the General Rules of Interpretation applied and written out, supporting rulings cited, alternatives ruled out on the record, timestamped throughout.
Entries cross-referenced against export and destruction records to surface drawback-eligible transactions, reclassification candidates and refunds from amended or vacated measures.
Tariff engineering, first-sale valuation, unbundling of non-dutiable cost, origin and substantial transformation, FTZ and bonded structures — each with its documentary burden stated.
Every action, annex amendment and exclusion list read the day it publishes and mapped onto your lines and in-transit shipments, with the deadline on the clock. Delivered as TariffOS.
Supplier networks reconstructed from customs, registry and shipping records so concentration, transhipment and restricted-entity adjacency are visible below tier one — before a detention makes them visible.
Built from the public record alone. Send a company name; nothing else is required, and there is no obligation on either side.
A continuously updated register of published trade actions, screened against the public trade record and your own shipments. Company-specific findings the same business day a measure lands — with the deadline that follows it running on the clock.
Open the live register →Language models are extraordinary at volume and unreliable at consequence. The pipeline is built around that asymmetry, and it is the same pipeline every time.
How we work →One scoping session with your trade, procurement and legal leads: which measures matter, which entities are in scope, and what a defensible answer has to look like to be usable.
Regulatory texts, annexes and binding rulings; manifest, registry and corporate records; your catalogue, bills of materials and entry history — normalised once, provenance retained per field.
Models read the annexes, reconcile naming and transliteration variants, follow derivative lists and draft the reasoning. A quarter of compliance-team work, processed in hours, uniformly.
Analysts review each flag that moves money or creates liability. Confidence is reported, disagreement is recorded, and nothing arrives as a conclusion when the record supports only a range.
Findings arrive ready for review by your licensed customs broker or trade counsel, who determine classification and make all filings. Where the measure keeps moving, we keep watching and reissue.
Official publications, binding rulings, public manifest and registry records, plus your own documents. Each flag carries the one it derives from.
Which rule was applied, which alternatives were considered and why they were ruled out — the record that demonstrates reasonable care.
Manifest data covers ocean cargo only. Invoice values are not public. Where the record cannot settle a question, the finding says which document would.
Client data is processed under engagement terms, never pooled across clients, never used to train third-party models. NDA before anything is sent.
A first briefing is built from public records alone — no documents required, no obligation. Findings are typically returned the same business day.