TariffWolf Logo TariffWolf™

Episode 4: HTS Classification at Scale: Auto-Classification, Audits and AI

Leaders Speak

Shawn McBrien
Speaker

Shawn McBrien

Founder - McBrien Customs Consulting

Shawn McBrien is a customs compliance and international trade professional with over 15 years of experience in import and export classification, regulatory compliance, and operational improvement. He currently serves as Manager of Import Compliance at Bealls, Inc. and is the founder of McBrien Customs Consulting. A certified Lean Six Sigma Black Belt and U.S. Army veteran, Shawn specializes in HTS classification, customs audits, process automation, and risk reduction, helping organizations strengthen compliance while improving operational efficiency.

TL;DR

Shawn McBrien, founder of McBrien Customs Consulting, has classified goods for automotive, high-tech, mining and retail companies. He explains how part logic and rule-based auto-classification handle large volumes, why every classification needs a trail of evidence, how to size a team with time studies, and what he found when testing AI classification tools. He also shares a case where one wrong engineer nearly changed a client's classification.

5 Key Findings

  1. Customs logic decides, not your instinct

    Shawn compares customs to the boy in The Matrix who says there is no spoon. CROSS rulings show how CBP actually reasons, so check the newest rulings and watch for revoked or modified ones.

  2. Automate the simple items, audit the output

    Part-number logic and rules such as men's cotton T-shirts let teams classify groups of ten to a hundred items at once. A risk-based sample audit confirms the rules work, and senior time goes to complex items.

  3. Leave breadcrumbs for every decision

    Record where each fact came from. In one case a client's engineer called a part off-the-shelf, but he did not work on that model. Shawn's files showed it was custom-molded for the client.

  4. Size the team with time studies

    Throughput ranges from six parts an hour at audit level to fifty at a service provider. Time several people on the same sets, ask for historical item mixes, and audit past classifications while you build the library.

  5. AI tools still invent codes

    In Shawn's tests, AI tools misread customs logic, relied on outdated rulings and returned HTS numbers that no longer exist, even after being told to use the current tariff only. Treat output like a new analyst's work.

Advice to the Professional

Shawn McBrien From Shawn McBrien

Classify for compliance, not for the lowest duty. Shawn’s example is a tabletop waterfall: it is not a decorative resin article in Chapter 39, because CBP rulings treat the pump as its essential character. If customs asked, could you defend your choice with documents?

Do not rush analysts to hit a number per hour. Give newer staff routine items to build confidence, have senior analysts handle project parts, and have two people classify the same item to surface disagreements. Keep evidence in more than one place.

Conclusion

Shawn’s method rests on research, evidence and redundancy. Automation and AI can speed up the routine work, but a defensible classification still depends on someone checking the source data and following customs logic.