Remote monitoring for machine builders: how Pomki did it
By Yannic, CTO, VlegelPublished on Updated on Reading time: 6 min

- Customer
- Pomki
- Location
- Belgium
- Sector
- Machine building (food machinery)
As a machine builder, you sell a machine, then it lives at the customer’s site. Without a connection, you only hear about it when the customer calls. With remote monitoring, the machine reports its own status: running or stopped, faults, and how it is being used. Pomki did just that for its apple presses.
Who is Pomki?
Pomki builds machines in Belgium that press fresh apple juice on the spot: in shops, at fruit growers’ premises and at events. Customers watch their juice being pressed, while the machine collects the waste itself. In Germany, Kaffee Anders distributes the machines to retailers (FreshPlaza).
What was the challenge?
A Pomki machine sits in a shop, far from its builder, not in a factory next to the engineers. To find out whether it was running, had a fault or how it was being used, someone had to visit or call.
Every machine builder with machines in the field faces that question: what happens to my machine when I’m not there?
What did we do?
- Connect. The machines were given a secure connection to the cloud (Microsoft Azure). They report their status and faults, wherever they are.
- Collect. Data from all machines comes together in one place, with the same name and meaning for each signal. That makes it possible to compare machines.
- Understand. In Power BI, the team can see how each machine is running, on a PC or a mobile.
What does remote monitoring offer machine builders?
- Faster support. You see the fault before the customer calls, or already know what’s wrong when they do.
- Fewer trips. Many problems can be solved with a phone call or instructions, rather than sending a technician out.
- Better machines. Your R&D team sees which faults actually recur and how customers use the machine. That’s better input than a list of complaints.
- A service you can sell. When you know your machines, you can offer service, maintenance or uptime contracts.
How do you get started as a machine builder?
- Choose one machine type and one question. For example: "Which fault causes the most call-outs?"
- Define the signals you need. Not everything from the PLC, just what answers that question: status, fault codes, counters.
- Give each signal one name and one meaning. Use the same ones on every machine of that type. The first machine then becomes the template for the whole fleet.
- Read first, control later. Start by reading data. If you ever need to change something remotely, always leave that to a person, with the customer’s permission, and never change safety functions.
- Buy the platform, build the connection. You don’t need to build your own cloud or data platform. What you commission is the connection to your PLC and your machine’s data model.
What does the law say about connected machines?
Two European regulations make connected machines the builder’s responsibility:
- The Machinery Regulation (EU) 2023/1230 replaces the Machinery Directive on 20 January 2027. It adds a requirement for "protection against corruption": a cyberattack must not compromise a machine’s safety functions (Pilz).
- The Cyber Resilience Act (EU) 2024/2847 applies to products with digital elements. Reporting obligations for actively exploited vulnerabilities have been in force since 11 September 2026, with the full rules applying from 11 December 2027 (Travers Smith).
For what this means for your specific machine, speak to your compliance adviser. One thing is clear: a connection designed today with a separate data layer, read-only access by default and a person making the decisions is a stronger foundation than an open port to the PLC.
Where does this lead?
When you measure your entire fleet using the same names and meanings, you can compare machines. AI can then spot anomalies before a machine fails and suggest a likely cause to the service technician. It reads from the data layer, never directly from the PLC, and makes no changes to a running machine itself. The foundation is always the same: connected machines with reliable data.
Conclusion
A machine at a customer’s site doesn’t have to be a black box. Pomki connected its apple presses to the cloud and now sees how they are running in Power BI. Start with one machine type and one question, give each signal one name, read first and leave decisions to a person.
Do you build machines and want to know how they run at your customers’ sites? We’ll look at one machine type together and tell you honestly what to measure first. Book a chat
Read more: For machine builders · Bespoke or off-the-shelf software? · Industrial AI
Frequently asked questions
What is remote machine monitoring?
Your machines send you their status, faults and counter readings over a secure connection. That lets you see how they are running remotely.
Do I need my own IoT platform?
No. You can buy platforms such as Microsoft Azure. What you commission is the connection to your machine and the data model.
Is remote monitoring secure?
Yes, if you design it properly: read through a separate data layer, keep access to the PLC closed, and leave every remote change to a person. From 2027, the Machinery Regulation and the Cyber Resilience Act require this too.
Can I change machine settings remotely?
Technically, yes. Only do so with the customer’s permission, have a person make the change, and never change safety functions.
What does it cost to get started?
That depends on your machine and your question. That’s why you start with one machine type and one question, then expand once it works.

