From raw to structured data

Relevant information for HS classification, essential in customs clearance, and thus the correct pricing of goods can come from a variety of sources. Packing slips, purchase invoices and transport and insurance documents in numerous formats; documents that are also shared or changed at different points in the chain. To recognize the classification of document types and to correctly interpret and automatically process certain content, we are developing algorithms together with Pharox from Venlo, TU Eindhoven and several companies within the work package E-Customs. The algorithms will be decisive in setting up the new platform.

Data are essential for building such a platform and collecting them proved to be a challenge during the Covid pandemic. With the data provided by shippers and distributors, we eventually managed to construct a first draft of an information portal. In the portal, Artificial Intelligence is used to transform unstructured data into structured data. Language processing and optical character recognition were further developed for this purpose. 
Pharox tested the technology at a customer's site on delivery notes. And with success: with the platform, time savings in the process can be realized of 20 to 35 percent.  The initial setup is now being further discussed with various logistics service providers for feedback and ease of use.  In addition to the technical developments, Pharox has approached several companies asking them to participate in a pilot. Both WSS and Seacon Logistics have agreed to participate. 

Partners theme E-Customs

Pharox, TU/e, Gerlach, Customs Support, Gaston Schul, ESC.