traide AI

Guest article by traide AI: The customs tariff: challenges for international online retailers and how traide AI simplifies customs tariff classification

Table of contents

The manual processing of customs formalities requires a considerable amount of time and carries the risk of errors. Customs tariffs in particular are causing difficulties for many retailers. The application of artificial intelligence (AI) can help overcome the challenges of pricing, which can help avoid financial losses and compliance issues.

What practical hurdles often arise in customs tariff classification? There are numerous challenges, such as outdated wording in the definitions/items of the customs tariff, numerous cross-references within the nomenclature or different interpretations of the customs tariff. In addition, the entirety of legal sources is often confusing and sometimes not accessible in different languages. It also happens that a large number of data records have to be correctly tariffed under time pressure, and missing product information and/or unclear goods descriptions make goods identification very time-consuming, sometimes even impossible.

How can artificial intelligence help with pricing?

Compared to fuzzy matching and decision tree solutions, deep learning technology and the Large Language Model (LLM) offer a comprehensive and unrestricted option for automatic product pricing. By using proprietary AI models and high-quality data sources, the AI can be trained to independently learn a function from the goods description and other input parameters for the tariff number. The training data, also known as ground truth data, should be the published classification decisions in the EU and the decisions and explanations in Switzerland in order to ensure objectivity. The AI is modeled by a huge neural network that distills knowledge from historical decisions from proprietary data sources into the neural network during the training process.

In contrast to pure memorization of data, AI learns to generalize the underlying concepts and rules and to understand the correlations. This enables the AI to achieve greater accuracy in pricing. Collaboration between AI software and human intelligence generally leads to higher productivity and lower error rates. AI can recognize errors that humans miss. Conversely, humans have knowledge that AI lacks. This “four-eyes principle” – or “human in the loop” – is particularly important in ambiguous or difficult pricing cases, for example when information is missing or the decisions are complex.

Better sooner than later: how can you save costs and hassle?

Correct tariff classification is of great importance. Errors in this area can have far-reaching consequences: Compliance and financial risks such as the loss of customs authorizations and simplifications, increased controls and customs audits, subsequent levies or overpayment of duties are just some of the possible consequences. To avoid such risks, retailers should already consider the customs tariff implications during the purchasing process and ensure that they use the correct tariff numbers.

About the author

Lilla Zsitnyánszky is Head of Customs at traide AI GmbH. The Berlin-based company traide uses AI technology and combines it with the support of long-standing customs experts to improve the security and effectiveness of customs tariff classification. Further information on the topic of technology-driven customs tariff classification can be found in the traide blog.

Experts from GJS will be happy to assist you with all questions relating to fiscal representation. You can
tobook a consultation appointment .

Do you want to expand your online shop to Switzerland and gain satisfied Swiss customers?
Arrange a free consultation!