Digitalization in logistics sounds like a seven-figure budget and a three-year program. New TMS, new portal, a consulting team in the meeting room. At the end, dispatch still retypes orders and calls drivers.
We ran two logistics projects in the last months. Both were in production in under four weeks. Neither replaced a single system. This piece shows what they have in common, where they broke, and how to pick your first process.
Technology is not the problem. Telematics delivers positions. The TMS knows the tours. The warehouse system has an import. The data is there.
What is missing is the piece in between. The service that compares a position with a time slot and notifies the dock. The service that reads a PDF and creates the order. Today a person fills those gaps, usually the person who already has the least time.
Strategy& and BVL measured this in November 2025 among 115 logistics providers in Germany, Austria and Switzerland. 96 percent have started. 76 percent automate processes. Only 10 percent scale new technology successfully, and a quarter do not measure results. Everyone experiments. Few get to production.
Three reasons the gaps stay:
A forwarder with 40 vehicles did not know when its trucks would reach the dock. The telematics showed every position and reported nothing. Dispatch called drivers. The company estimated around four hours a day for status questions across the team.
Today a service reads the telematics every 15 minutes. It calculates the arrival against the tour's time slot and notifies customer and dispatch 30 minutes before the dock. If the arrival shifts by more than 15 minutes, a second notice goes out.
Technically this first ran through the telematics interface because there was no open API. Later we switched to the system's position report. Dispatch never noticed. The full story is in the arrival notice case study.
Result: zero calls to drivers about arrivals. Dispatch puts the time into tour planning.
A wholesaler received 50 orders a day as PDF by mail. Eight customers, eight formats. One clerk retyped for three and a half hours a day. Five to eight percent of orders had errors.
Today a service reads the mailbox every five minutes. It recognizes the format by sender, reads lines and addresses, checks them against the article master and creates the order in the WMS. Unclear cases go to a list, PDF and proposal side by side. Details in the order capture case study.
Result: under an hour a day, almost all review cases. Error rate below one percent.
No system was replaced. Telematics, TMS and WMS stayed. The service sits between them. That is why it takes four weeks and not three years.
API when there is one. Screen when there is not. We use the interface as soon as it exists. If there is none, the service operates the system like an employee. Slower, but it runs, and it can be switched later.
People stay in the process, but only for decisions. The service books what is unambiguous. The clerk gets the unclear cases with everything needed to decide.
Operations are part of the delivery. Every run writes a protocol. A watchdog checks not whether the service ran, but whether it delivered: notices, bookings. Running without delivering is an outage and gets reported as one.
Dynamic interfaces. The telematics shows placeholders on first load. The real values arrive one to two seconds later. The service had to learn to wait.
Relative times. "In 2 hours 30 minutes" instead of a clock time. With time zones and daylight saving, two days lost.
Format variety. Eight PDF formats are eight reading rules. Two of them carried 70 percent of the volume. We could have started a week earlier.
Silent failures. A service can run and deliver nothing for days if a step before it fails quietly. Since then the watchdog counts bookings, not runs.
Three questions are enough:
The process with the most manual work and stable rules comes first. Only when it runs in production does the next one start.
Orders, portal, arrivals or reconciliation: which one costs your dispatch the most minutes a day? Name the one. I will show you, on our example, how it runs automated. Ask for a process check: 30 minutes, no sales pitch.
