Skip to content
Artificial Intelligence

AI brings more efficiency & greater accuracy to product classification

Suzanne Offerman  Senior Proposition Manager, ONESOURCE Global Trade

· 5 minute read

Suzanne Offerman  Senior Proposition Manager, ONESOURCE Global Trade

· 5 minute read

If your business imports or exports finished goods, components, or parts, you know firsthand the challenges of product classification.

First, cross-border trade requires a plethora of rules and standards — hence, the World Customs Organization (WCO) which includes more than 130 member countries and represents 183 customs administrations that process about 98% of world trade.

Next, the Harmonized Schedule (HS) provides guidelines for categorizing products and is the primary factor in determining which tariffs apply and the amount of duty to be paid when goods move across borders. Classification codes streamline the system: WCO uses 6-digits and countries have additional 2, 4, 6 or even 10 additional digits.

This would be vastly complex to manage even if circumstances were static — which, of course, they’re not. Regulatory updates to the tariff schedule can occur 10 or 15 times or far more each year. In 2018, for example, the U.S. updated the tariff schedule 270 times and Brazil updated its schedule 244 times. Meanwhile, growing protectionism and tariff wars add to the uncertainty and challenge of managing supply chains.

Companies also have to consider their own impact on classification requirements — to ensure, for example, that proper classification occurs when a new product is created or an existing one is modified. “The proper classification of a product drives so much,” said Virginia Thompson, senior product manager at Thomson Reuters, during the recent webinar, AI at Your Fingertips: How Emerging Technology Is Redefining Classification.

“It drives the revenue that a country is going to be realizing and that has become particularly important in recent years as we’ve seen increasing tariffs and… retaliatory tariffs,” Thompson explained. “It drives the revenue tied to things like preferential claims (and) anti-dumping and countervailing duties.”


You can learn more from two recent webinars, now available on demand, AI at Your Fingertips: How Emerging Technology Is Redefining Classification and The Impact of COVID-19 on the Global Supply Chain.


As a result, customs enforcement is rigorous, and the stakes are high for companies. “It’s very important to a company that they get their classification correct,” Thompson noted. “That’s why the documentation and the proof to Customs when they come in for an audit is very important — that you’re able to prove to Customs that you have classified your products correctly.”

Reasons for classifications

There are three crucial reasons businesses must correctly classify goods: compliance, cost, and customers.

Compliance — Multinational companies often find departments or businesses within their organization that independently track and execute classification changes with little or incorrect communication across the business and to customers. This siloed approach can lead to discrepancies that result in shipment delays, incorrect duty payments, and compliance violations. 

Cost — Sometimes, what you don’t know can hurt you very much. Being unaware of classification changes carries a financial cost — often, overpayment or underpayment of duties. Companies may be unaware they have overpaid and, if the mistake is realized, the process to recover refunds can be complex and lengthy. An underpayment of duties, on the other hand, can result in fines as well as delays because the correct duty payment is required at the time of shipment.

Customers — Shipment delays and supply chain disruption can affect both the company and its customers, so correct application of HS classifications is paramount.

Many businesses rely on global trade management (GTM) software or tariff classification solutions to help classify products correctly. This type of automation can deliver:

  • accuracy through repeatable, documented processes;
  • cost savings because classification occurs faster and avoids mistakes that increase risk of non-compliance; and
  • more time for team members to focus on strategic work, because they spend less time on manual tasks.

The complexity of classification, including how often changes are required, makes it difficult to be completely automated — however, artificial intelligence is closing the gap.

During the webinar, Randy Rotchin, president and CEO of commodity classification company 3CE Technologies, called the involvement of AI in classification Smart HS. “Smart HS essentially is a tool that uses artificial intelligence to assist with classification,” Rotchin said. “Although it looks and acts like a search engine, it really is a decision support system that’s built upon the expert’s model of the domain, and that’s what differentiates it from other classification wizards or… tools.”

AI-based solutions read and interpret readily available, complex commercial product information, Rotchin explained, and bridge the gap between the way products are described by the trade and how they are expressed in the Harmonized System.

“To give you an example of how wide that gap could be, if you didn’t have explicit knowledge that baby food is expressed in tariff schedules as ‘homogenized composite food preparations’ you would be knee-deep in that gap,” he said, adding that AI classification systems use “an inference engine to assess the completeness of goods descriptions according to HS rules.”

Then when a product description is submitted to the system it analyzes the goods description to determine whether or not there are adequate details according to HS requirements to assign a code directly, he said.

More answers