Tariff numbers found within seconds – tariff classification AI-powered

If you want to sell abroad, there’s a lot to consider. Especially if there is a customs border between where the goods are stored and the customer's shipping address. For import and export, each item needs a customs tariff number based on its properties. Because without it, nothing works at customs! However, the determination is a manual, time-consuming process in many companies.

As a specialist in cross-border logistics, Beyond Borders – a joint venture by MS Direct and Seven Senders – clears the way for millions of articles to enter markets in third countries. To manage the tariff determination for all these goods, we rely not only on the knowledge of our customs experts, but also artificial intelligence (AI) using machine learning – our TariffTranslator. The automated determination of the customs tariff number from master data significantly simplifies customs processing so that goods can cross the border smoothly, within the shortest possible time.

How exactly does this work, you ask? Eva Tyssen, our Head of Business Development and Customer Success, got answers from asked Prof. Dr. Siegfried Handschuh from the University of St. Gallen (HSG), who, with his team, helped to develop our AI tool.

"As a company, you need to be aware that you are sitting on a treasure with your process data."

Eva: Siegfried, we approached you almost two years ago with the aim to automate our tariff number classification. How quickly did you realize that this was a perfect use case for machine learning?

Siegfried: We had an idea, but we didn't know it until we worked together on a proof of concept and were able to use real-life sample data. It quickly became clear that we could achieve very good results with machine learning. The available amount of high-quality data makes this use case ideal.

Eva: How did we train the machine? What did we need to do and how long did the process take?

Siegfried: The project took about three months, of which about one month was spent on data handling, one month on working with the training algorithms, and one month on the user interface. Such projects always require extensive data exploration and preparation. The preparation of the data – we had about 57 million records – included a quality check, an examination of patterns, linking of data sources, and detection of anomalies. With the data prepared in this way, we developed a total of 75 machine learning classifiers for the task. These models are trained, validated and fine-tuned. In addition, we have created an infrastructure that allows for subsequent post-training, making the solution open to the future.

Eva: What do you generally recommend to companies that want to optimize processes with AI?

Siegfried: One is to check whether the processes can even be addressed with AI, i.e. what does the current solution look like in terms of costs, error rates and process inefficiencies, and and the expectations of the AI solution? Secondly, AI needs data. As a company, you have to recognize that you are sitting on a treasure of process data that needs to be mined. When you are aware of that, you will easily find use cases to automate business activities and processes.

Eva: Thank you very much, Siegfried, for this interview!

Any questions? As a leading partner for e-commerce cross-border logistics, we support you not only with customs clearance and shipping but also with fulfilment and returns management. Just get in touch with us! Eva is looking forward to your message:

eva.tyssen@ms-direct.ch