Three categories of AI platforms for transport in the Netherlands
The market for AI in transport can broadly be divided into three layers. First layer: international SaaS platforms that bundle route planning, fleet management, and track-and-trace into a single subscription. They deploy quickly, are reasonably priced, and offer enough functionality for a mid-sized transport company with straightforward operations. Second layer: building an AI layer on top of an existing TMS or WMS. Think automating order processing from email, matching freight documents to trips, or handling driver communication through a chat interface. You keep your current system but add targeted automation at the points where people spend the most time. Third layer: fully custom software. A new core system, built around the processes that set you apart, with AI embedded from day one. Not every company needs this, but for businesses with complex tariff structures, proprietary planning logic, or specific customer requirements, it is the only approach that truly scales.
Where a generic AI platform for transport in the Netherlands falls short
The Dutch transport market has a number of characteristics that international platforms consistently underestimate. Cabotage rules determine how many trips a foreign truck may make within the Netherlands: generic platforms rarely enforce these correctly. The truck levy, introduced in 2026, requires per-vehicle mileage registration and integration with tax reporting. eFTI, the European digital freight document framework, is gradually replacing paper CMR documents, but requires a data model that genuinely supports this. Driver communication in the Netherlands runs through a mix of apps, portals, and standard messaging similar to WhatsApp, depending on the customer. A platform that does not offer this as a configurable component means dispatchers still end up switching between systems manually. These are not edge cases. This is daily operations.
What an AI layer on an existing TMS or WMS concretely solves
Many transport companies in the Netherlands run on a TMS purchased five or ten years ago. The system works, but data entry is manual: orders arrive by email, someone re-enters them, freight documents are scanned and stored without further processing. An AI layer addresses exactly those processes. Order processing from unstructured emails, including routing to the right planner. Automatic extraction of CMR documents, packing lists, and customs paperwork. Alerts for deviations in route planning or weight. These are targeted AI Workers running alongside the existing system, without requiring you to replace the core platform. The fit is strong when the TMS itself still holds up and the bottleneck is the manual work surrounding it.
When do you choose fully custom software?
Fully custom software is the right choice when the core system itself is blocking growth. Three signals point to this. First: your tariff logic or planning rules are complex enough that no standard package supports them without significant workarounds. Second: data lives in spreadsheets or standalone files because the TMS cannot handle the data model. Third: every integration with a customer or client requires a custom connection because the system has no flexible API. In these cases, an AI layer helps in the short term but does not resolve the structural problem. The smarter move is to rebuild the core system from scratch, AI-native from the ground up, resulting in a system you own that follows your processes precisely. This takes more time than a SaaS subscription, but it produces a system no one else has and one that does not get more expensive every year.
Decision framework: which approach fits your transport operation?
Use this as a starting point. Choose a SaaS platform if your operation is relatively standard, you want to move quickly, and Dutch specifics such as cabotage, eFTI, and the truck levy are not a significant part of your daily work. Choose an AI layer on your existing system if the TMS itself still functions well but the manual work surrounding it costs too much time. Think order processing, document flows, and driver communication. Choose fully custom software if your planning logic, tariff structure, or data model is specific enough that standard packages consistently fall short. The most important question is not which platform has the most features. The question is: which processes in your operation are distinctive enough that forcing them into a generic system is not an option?