Why order intake in particular
Three reasons. It is high volume: every order passes through the same hands. It is repetitive: nobody grows from it, least of all your best people. And it is measurable: you know exactly how many orders come in per day and how many minutes the retyping costs. That keeps the business case simple, and the result shows up within weeks rather than quarters.
What happens technically
Incoming orders are picked up from the mailbox, whether they arrive as PDF, Excel or plain text. The AI reads the content in context: collection and delivery address, date, load, references. There are no per-customer templates, because those fall over the moment a customer changes their format. The data is validated and entered into the TMS as a draft order; the planner checks and confirms it.
A person stays in the loop, deliberately
Operational transport is full of exceptions: an ADR load, an unusual time window, a customer who means something other than what they write. Fully autonomous processing is not the goal here. The AI does the typing and flags the doubtful cases; the planner decides. That is how you win back the hours without losing control.
What it delivers
The quiet hours spent retyping vanish, and so do the errors that come with manual work. Something else emerges that is often underrated: structured order data from minute one. With it, margin per trip, CO2 reporting and invoicing suddenly become achievable, because they all hang off the same data that is now already correct at the source.
From 1 July 2026 the truck toll and eCMR add one more reason: trip and document data has to be available in structured form anyway. Automate the order intake and you get that foundation thrown in for free.