Automatically Apply Tariff Codes to Imports and Exports
A “Big 4” Accounting firm needed to correctly tag imports with US tariff codes. Differences between manufacturers’ product descriptions and US Customs descriptions lead to applying the wrong tariff.
Incorrect Tariff Codes can lead to higher taxes and even fines. Generally, this process involves a skilled labor-intensive manual review, which is costly and time consuming. Standard machine learning techniques act as a blackbox which provides the decision but not the explanation.
The Lymba Solution
Use deep learning, NLP, and semantic inference to process product descriptions and United States Customs descriptions (HTS rules).
Process
First, we created a custom domain ontology to cover the commonsense product information and bridge the gap in language between product description and HTS.
Output
Then, we output a report with recommended HTS codes (Tariff schedule) and rationale for each product being imported, allowing for faster review, better accuracy, and lower costs.