Facebook IconFacebook IconFacebook Icon

Automated Customs Code Attribution

Sagacify Logo

Challenge

Jan De Nul is a company specializing in building complex maritime infrastructures such as offshore energy production stations, preservation of waterway depths projects, construction of ports, etc. Its logistics department is constantly busy transporting materials and spare parts internationally. This creates a lot of additional administrative tasks. One of these tasks is that every shipped package needs to receive a specific Harmonised System (HS) code that determines its taxation rate. The attribution of these customs codes is a very complex task since the codes not only depend on the product type, but also on the country the package is imported into.

Since there are fine nuances between customs codes, choosing the right code for a specific product is complex, and any mistake has an important impact. Mistakes could result in paying an overpriced tax, a regularization fine, or blocked packages at customs, which is especially problematic due to extended delays for customers.

Because of the complexity of the task, finding the right custom code requires the collaboration of an international customs officer and local customs agents. 

To avoid money and time losses, Jan De Nul was looking for an efficient solution to identify the right customs codes for its shipped goods.

Solution

Jan De Nul has a catalog of products with their technical description and internal code. However, the descriptions of the products in the customs catalog are very different from the ones of the company because they do not use the same language register. The goal was therefore to match the customs codes with the internal codes. To this end, the solution developed by Sagacify consisted of 2 models called the retriever and the reader:

  • Retriever: The retriever was responsible for finding the related context. This acted as a filter and allowed the retrieval of all possible section/subsection candidates. At this step, it doesn't matter if we retrieve too many candidates because we just want to be sure not to miss any. 
  • Reader: The reader selected the best candidate. Therefore, it predicted the section/subsection for a given line item from a given context (provided by the retriever).

Key element  

  • The combined work of two models allows selecting the best potential HS code for a specific product
  • The solution takes into account the cost of an error and chooses to automate the decision, or not, depending on the risk of automating the decision

Results

Our solution was able to process any texts, images, or both, and automatically attribute codes or labels to the presented data element.

It brought an efficient way of managing customs codes by attributing the codes faster, but also more accurately than the control dataset.

It also reduced helped reduce costs, and prevent the company from fines.

Related cases