Freehand Studio · AI Agent · Trade & Compliance

HS Classification Agent: Every Product Correctly Classified Across Every Active Trading Jurisdiction

Products classified to the correct HS code using product descriptions, attributes, and material composition validated against ruling databases and jurisdiction-specific tariff schedules. Freehand's HS Classification Agent delivers 95%+ accuracy at catalog scale.

Shipper
3PL
LSP
Carrier
Service Provider
95%+
Classification accuracy vs. 60€“70% typical with manual specialist review
1-3%
Of total import duty spend recoverable through correct classification alone
Weeks
To classify a full product catalog vs. months of manual trade specialist effort
Trusted by global leaders in logistics, manufacturing, and retail
Awards and Recognitions
The Problem

Manual HS Classification Is Error-Prone, Slow, and Doesn't Scale at Product Volume.

HS classification is complex, jurisdiction-specific, and changes frequently. Manual classification is error-prone and slow. Misclassified products result in incorrect duties, customs delays, compliance penalties, and post-shipment audit exposure.

Classification Complexity Exceeds Manual Capacity

HS codes are jurisdiction-specific, updated annually, and require interpretation of product attributes, material composition, and end-use context. Manual classification takes two to five minutes per SKU. At catalog scale, full coverage is not achievable within normal timelines.

Misclassification Discovered at Customs

Classification errors surface when goods arrive at the border. By then, the shipment may face delays, duty reassessment, and potential penalties. Post-entry corrections require formal protest processes that are time-consuming and not always successful.

Coverage Gaps Across the Catalog

When classification requires specialist time, only a fraction of the catalog is classified correctly. New products launch with provisional codes. Large portions of the item master carry classification risk that is invisible until customs flags it.

Reclassification After Schedule Changes Not Systematic

Tariff schedules change. Products that were correctly classified become misclassified after a schedule update if reclassification is not triggered automatically. Manual processes have no mechanism for monitoring schedule changes and triggering reclassification across affected items.

Audit Exposure Without Version History

Manual classification records are inconsistent. When a customs authority audits classification decisions, the documentation required to support the codes applied product attributes reviewed, rulings referenced may not exist in a format acceptable for audit defense.

Classification Inconsistent Across Trading Countries

A product traded across 22 countries may carry different HS codes in different jurisdictions correctly, because schedules differ but managing this complexity manually creates gaps where some country codes are correct and others are estimated or outdated.

What the Agent Does

Classify at Scale. Validate Against Rulings. Version Every Decision. Apply Automatically.

Classifies products to the correct HS code using product descriptions, attributes, and material composition validated against ruling databases and jurisdiction-specific tariff schedules. Classifications versioned, auditable, and auto-applied to shipment records.

Multi-Attribute Product Classification

Products classified using description, material composition, end-use function, and manufacturing method. Classification engine validated against tariff schedule databases from Descartes and Thomson Reuters for each trading jurisdiction. Ruling database references cited in the classification record.

Full Catalog Coverage

Entire item master classified not just high-value or high-volume SKUs. New products classified at item master creation. Seasonal and promotional products classified before activation. Coverage gaps eliminated from the first deployment cycle.

Automated Reclassification on Schedule Changes

Edge cases products with ambiguous classification, new material categories, or rulings under challenge routed to the compliance review queue for trade specialist review. Specialists handle exceptions; the agent handles volume.

Versioned, Auditable Classification Records

HS codes managed per trading jurisdiction. Products classified correctly for each country's specific tariff schedule. Jurisdiction-specific classification applied to shipment records automatically at booking.

Jurisdiction-Specific Classification Management

Pass all checks and route to the audit queue automatically. No manual review, no re-keying. Audit agents receive clean, normalized documents.

Exception Review Queue

Edge cases products with ambiguous classification, new material categories, or rulings under challenge routed to the compliance review queue for trade specialist review. Specialists handle exceptions; the agent handles volume.

Agent Handoffs

From Product Master to Compliant Shipment Records

Receives contract and product data from the Contract Ingestion Agent. Delivers classification data to the Duty Calculation, FTA Qualification, and Compliance Screening agents.

Receives from

Contract Ingestion Agent

  • Product specifications, material composition descriptions, and trade-relevant product attributes extracted by the Contract Ingestion Agent from supplier contracts and product documentation used as classification inputs.

This Agent

HS Classification Agent

  • Classifies products to the correct HS code using product descriptions, attributes, and material composition. Validates against ruling databases and jurisdiction-specific tariff schedules. Versions every decision. Applies classifications to shipment records automatically.

Triggers

Duty Calculation Agent

  • Verified HS codes delivered to the Duty Calculation Agent as the primary input for pre-shipment duty estimation and entry duty calculation.

FTA Qualification Agent

  • Confirmed HS codes delivered to the FTA Qualification Agent as the starting point for rules-of-origin analysis and free trade agreement eligibility determination.

Compliance Screening Agent

  • Product HS classifications delivered to the Compliance Screening Agent for restricted party screening and export control classification cross-referencing.
Before AI → After AI

What Changes When HS Classification Runs on the Agent

The product catalog does not change. The classification coverage and accuracy across it does.

Before the Agent
With HS Classification Agent
HS classification done manually by trade compliance staff. Two to five minutes per SKU. At catalog scale, full coverage is not achievable within normal launch or sourcing timelines.
Products classified automatically at item master creation. Full catalog coverage from the first deployment cycle. Compliance staff reviews exceptions only.
Misclassifications discovered at customs. Shipments face delays, duty reassessment, and potential penalties. Post-entry corrections require formal protest processes.
Classification accuracy improves to 95%+. Customs delays from misclassification reduced significantly. Duty exposure from incorrect codes identified and corrected before shipments move.
New products launch with provisional or estimated codes. Coverage gaps across the item master carry classification risk invisible until customs flags it.
New products classified at item master creation. Every product carries a current, jurisdiction-specific HS code before it enters any shipment or trade document.
Tariff schedule changes leave some products on outdated codes. The misclassification is discovered at customs, not at schedule change time.
Tariff schedule monitoring active for all jurisdictions. Changes trigger reclassification automatically. No product is left on an outdated code after a schedule update.
Manual classification records are inconsistent. When customs audits classification decisions, the documentation required for defense may not exist in an acceptable format.
Every classification decision versioned and logged with product attributes, tariff schedule references, and ruling citations. Audit trail complete and immediately available for customs review.
Measured Outcomes

Results from Live Deployments

Outcomes measured from consumer electronics, automotive, industrial, and pharmaceutical enterprise deployments across global trade operations.

95%+
Classification accuracy vs. 60€“70% typical with manual specialist review
1-3%
Of total import duty spend recoverable through correct classification alone
Weeks
To classify a full product catalog vs. months of manual trade specialist effort

Full catalog classified not just high-value or high-volume SKUs.

Classifications validated against ruling databases and jurisdiction-specific tariff schedules.

Automatic reclassification triggered when tariff schedules change.

Every decision versioned and logged. Complete audit trail available for customs authority review.

Connects to ERP product master, Descartes, Thomson Reuters, and TMS systems on day one.

Scales with catalog size. No incremental trade specialist headcount as product count grows.

Integrations

Works Where Your Product and Tariff Data Already Lives

Reads from ERP product master and tariff content providers. Writes classified HS codes to shipment records, trade platforms, and downstream compliance agents.

ERP

SAP S/4HANA · Oracle Fusion · Dynamics 365

Product master and item attribute data consumed via BAPI and OData. Description, material composition, end-use function, and manufacturing method pulled for classification.

Tariff Data

Descartes · Thomson Reuters

Jurisdiction-specific tariff schedule databases and ruling databases consumed via REST API for classification validation and schedule change monitoring.

Trade Content

QAD GTTE · Thomson Reuters

Trade content providers consulted for ruling database references and country-specific classification guidance.

Classification History

Freehand · Trade Platform

Prior classification history and binding rulings from Freehand and connected trade platforms used for classification consistency and ruling database cross-reference.

Middleware

MuleSoft · Dell Boomi · Seeburger BIS

Product and classification data flowing through your integration layer accessed without trade platform pipeline disruption.

PLM

Product Lifecycle Management Systems

Product specification and material composition data from PLM systems consumed via REST for classification at item master creation.

Product Master

ERP · Trade Platform

Classified HS codes written to the product classification master in ERP and connected trade platforms via BAPI and REST.

Shipment Records

SAP TM · Blue Yonder · Oracle OTM

HS codes applied to shipment records at booking in TMS systems via REST. Every shipment carries the correct code before it moves.

Downstream Agents

Duty Calculation Agent · FTA Qualification Agent

Verified HS classifications delivered to the Duty Calculation Agent and FTA Qualification Agent as primary inputs for duty estimation and trade agreement eligibility analysis.

Compliance Queue

Compliance Review Queue

Edge cases and ambiguous classifications routed to the Freehand compliance review queue for trade specialist review before activation.

Data Lake

Snowflake / Databricks

Classification records and audit trail data written to your data lake for compliance reporting and customs authority documentation.

SharePoint

SharePoint / OneDrive

Classification audit trail and compliance documentation stored to SharePoint for trade compliance archive and customs audit support.

14,000
Active SKUs classified across 22 trading countries in under 3 weeks in electronics deployment
$2.3M
Annual duty savings from correct classification in that deployment
95%+
Classification accuracy vs. 60% manual coverage prior to deployment
Day 1
Connected to ERP product master, tariff databases, and TMS shipment records from go-live
Case Studies

14,000 SKUs. 22 Countries. 3 Weeks. $2.3M in Duty Savings.

Real outcomes from enterprises running the HS Classification Agent in production.

Case Study 01

Global Consumer Electronics Manufacturer

Consumer electronics manufacturer with 14,000 active SKUs traded across 22 countries. Prior manual classification coverage was 60% of the catalog. The remaining 40% carried provisional or estimated codes.

Consumer Electronics · 14,000 SKUs · 22 Trading Countries

14,000 SKUs

Classified in under 3 weeks

$2.3M

Annual duty savings from correct classification

  • 14,000 active SKUs classified across 22 trading countries in under 3 weeks replacing a manual process that had covered only 60% of the catalog
  • $2.3M in annual duty savings identified from correct classification of previously misclassified or unclassified products
  • Full audit trail for every classification decision available from the first deployment week, replacing the inconsistent manual records that had created customs audit exposure
Case Study 02

Global Pharma Manufacturer

Pharmaceutical manufacturer with frequent product launches requiring HS classification in 15 trading countries. Manual classification was creating launch delays as trade specialist availability bottlenecked the process.

Pharmaceutical · 15 Trading Countries · Launch Delay Risk

Zero

Product launch delays from classification bottlenecks after deployment

95%+

Classification accuracy vs. 68% manual specialist baseline

  • Classification at item master creation eliminated the trade specialist bottleneck that had been delaying product launches by two to three weeks per market
  • Classification accuracy improved from 68% to 95%+, reducing customs examination events and duty reassessments across the global product portfolio
  • Automatic reclassification on tariff schedule changes ensured no product carried outdated codes after annual schedule updates in any of the 15 trading countries
Technology

Powered by the Freehand Context Graph

Classification accuracy requires product context, tariff schedule knowledge, and ruling precedent all in one place.

The Context Graph connects ERP product master data, tariff schedule databases from Descartes and Thomson Reuters, prior classification history, and ruling database references into the unified classification context. Every classification decision draws from the complete set of relevant data across jurisdictions.

Built on the Freehand Logistics Language Model, trained on Harmonized System classification frameworks, product attribute analysis methodologies, and ruling database patterns across global trade jurisdictions. It understands how product descriptions map to HS codes across different country schedules.

  • Every classification decision is versioned and traceable. The product attributes reviewed, the tariff schedule consulted, the ruling references cited, and the classification outcome are all logged. Complete documentation available for customs authority audit.
  • The Context Graph learns from ruling outcomes and classification corrections. Classification decisions challenged and upheld or revised update the logic. Accuracy improves with every product lifecycle and customs interaction.
  • Classification intelligence flows into every downstream trade compliance agent. The Duty Calculation Agent receives verified HS codes. The FTA Qualification Agent uses classifications for rules-of-origin analysis. The Compliance Screening Agent uses them for export control cross-referencing.
Architecture Overview
DATA LAYER AI TEAM Contracted Rates Carrier Invoices Shipment Events EDI Feeds ERP Exports Rate Cards CG Context Graph Freehand LLM Unified Semantic Layer Domain-Specific AI Self-Learning Model IA Invoice Audit Agent 100% invoice coverage GL GL Coding Agent GL posting & allocation AF Accrual & Forecast Agent Live spend accruals SI Spend Intelligence Agent Finance-grade data ERP OUTPUT SAP · Oracle Cloud · Oracle JDE · NetSuite · via API & EDI
FAQ

HS Classification: Questions Trade and Finance Leaders Ask

Straight answers to what trade compliance and finance leaders ask before deploying the HS Classification Agent.

What product data does the agent use for classification?
+

Product description, material composition, end-use function, manufacturing method, and any binding rulings from prior classification history. All attributes pulled from the ERP product master and PLM systems at classification time.

How does the agent handle jurisdiction-specific classification differences?
+

HS codes managed per trading jurisdiction against country-specific tariff schedules from Descartes and Thomson Reuters. A product traded across 10 countries carries 10 jurisdiction-specific codes correctly different where schedules differ.

What triggers reclassification when tariff schedules change?
+

Tariff schedule monitoring runs continuously for all active trading jurisdictions. When a schedule update affects a classified product, reclassification is triggered automatically. The classification record is updated and the change is logged with the effective date.

How does the agent handle products that are difficult to classify?
+

Products with ambiguous classification new material categories, multi-function products, products with contested rulings are routed to the compliance review queue for trade specialist review. The specialist's determination is logged and applied to similar products going forward.

How does the HS Classification Agent fit into the Freehand pipeline?
+

Receives product specifications from the Contract Ingestion Agent. Delivers verified HS codes to the Duty Calculation Agent, FTA Qualification Agent, and Compliance Screening Agent as primary inputs for downstream trade compliance workflows.

How quickly can the HS Classification Agent be deployed?
+

Deployable in days via pre-built connectors to ERP product master and tariff content providers. Most enterprises have initial catalog classifications within the first week and full coverage within the first month of deployment.

Get Started

Deploy the HS Classification Agent Across Your Product Catalog

Full catalog classified. 95%+ accuracy. Audit trail complete. Deployable in days.

Built on Freehand Studio · freehand.ai

See how Freehand recovers margin you're already losing

Map your commercial agreements to real-world execution - recovering 2-5% in lost margins and ensuring 100% audit coverage.

What to expect in the call

We identify exactly where you’re leaking margins

See how our AI Teams cross-check contracts, and resolve overcharges

Get a savings estimate based on your current spend and systems.

Trusted & Recognized by

KEARNEY
pwc
Gartner

See AI teams in action