Under Active Development — This site and its services are being built incrementally. Features shown may be in preview or coming soon.
AI & Tariff Classification

AI-Enhanced, Expert-Driven Classification

We believe AI is most powerful when guided by deep customs knowledge. Our approach combines machine learning with expert judgment to deliver classifications that are not just fast, but defensible — grounded in regulatory understanding, not pattern-matching alone.

Strengths

Real Operational Gains

Speed and Scalability

Process thousands of products in seconds with up to 71% time savings per item. Manage catalogues from millions to over 100 million SKUs without proportional cost increases.

Consistency and Cost Optimisation

ML models trained on historical customs data ensure consistent code application. Advanced tools can model the total tariff stack across hundreds of origin scenarios, identifying duty reductions of up to 50%.

Audit Readiness

Modern AI platforms generate detailed rationales for classifications, citing relevant GRI rules and tariff notes, strengthening audit defence and regulatory transparency.

Limitations

Why Human Expertise Remains Critical

Lack of Legal Judgment

AI functions primarily as a pattern matcher. It struggles with nuanced application of GRIs and may fail to interpret legal precedents or complex product descriptions. It excels at "what" (data processing) but lacks capacity for "why" (legal reasoning).

Regulatory Volatility

Trade regulations can shift overnight. AI models trained on historical data may not immediately integrate changes. General-purpose models average ~72% accuracy, while purpose-built platforms can exceed 95–99% — but none is immune to regulatory lag.

Accountability and Confirmation Bias

The importer of record remains legally responsible under 19 U.S.C. § 1484. An AI confidence score does not constitute valid defence during audit. AI may reinforce pre-existing bias toward incorrect codes when prompted poorly.

The Industry Consensus: Hybrid Implementation

Leading practitioners recommend a human-in-the-loop model that combines AI speed with expert validation.

Step 1

AI as Research Tool

Surface multiple candidate codes, calculate duty impacts, perform document extraction

Step 2

Human Final Mile

Licensed customs brokers validate suggestions against current regulations, GRI logic, and product attributes

Step 3

Continuous Improvement

Feed validated decisions back into the system to improve model accuracy over time

Coming Soon

SKU-Level Compliance Management

A complete product compliance lifecycle — from initial classification through to ongoing assurance.

Import/Export Code Management

Maintain HS codes, Schedule B codes, and ECCN classifications at the individual SKU level with full audit history.

Certificate Tracking

Link certificates of origin, phytosanitary certificates, and other trade documents directly to each product SKU.

Condition Compliance Assurance

Monitor tariff conditions, preferential rate eligibility, and anti-dumping requirements with automated alerts when status changes.

Regulatory Change Tracking

Flag affected SKUs when HS nomenclature updates, FTA rule changes, or new trade measures are published.

"We use AI to accelerate research and eliminate administrative burdens, freeing our experts to focus on the complex decisions that protect your compliance and your bottom line."

— BCH Customs positioning