What is changing in customs declarations in 2026?
Two developments are putting additional pressure on customs operations. First, the EU is abolishing the import duty exemption for goods valued at 150 euros or less imported from outside the EU as of 1 July 2026. In its place comes a temporary flat fee of 3 euros per HS-code category within a package. That may sound manageable, but for parties handling large cross-border volumes it means every shipment must now be correctly classified. A package containing two product categories incurs a six-euro charge. Misclassification costs more. Second, Dutch Customs is intensifying its scanning and analysis capacity in 2026. Investment is being made in a central scanning environment that lays the groundwork for full digitalisation of declaration processing. Faster and more accurate controls mean that discrepancies in declaration data will be identified sooner.
Where does the customs operation get stuck in practice?
The pattern we keep encountering is this: the data required for a correct declaration is scattered across multiple systems. A purchase order lives in the ERP. The goods description comes from the supplier, in a PDF or email. The HS-code is looked up manually or copied from a previous shipment, even when the article has changed. The customs value must be calculated from invoice data that has not yet been verified. Each of those steps takes time and introduces errors. And with the increased control pressure from Customs, as indicated in the 2026 annual plan, those errors are no longer acceptable. Customs Support Group puts it plainly: customs has become a strategic function. Organisations that do not invest in digitalisation and data risk sanctions, disruptions, and fines.
Is this a software problem or a process problem?
Both, but in a specific order. The process problem is that customs data has no owner in the chain. Procurement populates the ERP. Logistics dispatches the shipment. The customs broker prepares the declaration based on whatever is provided, sometimes by email, sometimes through a portal, sometimes by phone. No one has a complete, current picture. The software problem is that the systems used by each of those parties are not designed to exchange that data in a structured way. Generic ERP packages contain customs fields that remain empty in practice. Classification tools operate independently from declaration software. Interfaces exist on paper but are bypassed in practice because they are too slow or fail too often. The result is a customs employee who manually hunts down, checks, and enters data every single day.
What helps and what does not?
Hiring an extra employee is the reflexive response, but it does not solve the structural problem. Demand for customs professionals in 2026 exceeds supply. Outsourcing to a customs broker is an option, but they need the same data from you as the instructing party. What does help is structuring data earlier in the chain. In practical terms, that means the HS-code is recorded when a purchase order is created, not after the fact. Supplier documents are read and validated automatically rather than retyped manually. And the connection between trade documents and declaration systems is direct, not routed through a stack of emails. That is precisely what document AI tooling does in the customs context: it imposes structure on unstructured input so that the employee reviews and approves rather than rekeys and searches.
When is a custom-built system the right choice?
Not every customs party benefits from building a new core system. When volumes are low and shipments show little variation, a well-configured standard package with a document AI layer on top is often sufficient. But for freight forwarders and importers processing hundreds or thousands of lines daily, with diverse suppliers, multiple countries of origin, and shifting tariff codes, a generic package falls short. The reason is always the same: the package was not designed for your specific supply chain. The classification logic, validation rules, integration with your own ERP, and reporting to Customs do not fit out of the box. Adapting costs more than building from scratch. In those cases, rebuilding the core system from the ground up, domain-specific and with customs logic built in from the start, is the better investment.