Knowledge · Machine builders
How can a machine manufacturer scale service without hiring more staff?
By Gerd Bart, Founder and Managing Director · Updated
The short answer
By having your team handle only the cases that need an expert: your customer finds documents and the right spare part on their own, whatever they can't solve reaches you complete, and the knowledge of your experienced people is available to everyone. Hiring more is rarely the answer, because skilled people are scarce and every new hire first has to learn what the experienced ones carry in their heads. With Transaction-Network, for example, your customer orders your parts from the record of their machine, sends Service Requests straight to your team, and the Service Brain answers questions from the documents you have released for it.
Why hiring more is rarely the answer
Every machine you sell grows your installed base, and with it the work in service: requests, spare parts, maintenance, faults. Your team doesn't grow at the same pace. The U.S. Bureau of Labor Statistics projects 14% employment growth for industrial machinery mechanics, with about 51,900 openings a year (Occupational Outlook Handbook, 2025). In German machine building, about 296,000 employees retire by 2034 and about 118,000 follow them, a gap of about 178,000 skilled workers; one in four employees is already 55 or older (IW for the VDMA's Impuls Foundation, November 2024).
And a new hire doesn't solve the knowledge problem right away: they don't know your older product lines and ask the people who do.
Where your service team's time goes today
Some requests don't need an expert. Your customer asks for a manual that already exists, for a part number, for price and delivery time, or for the status of their request. Then there are the questions your team has to ask: Which machine exactly, which serial number, which software version?
No study tells you how large that share is for you. Count it yourself for a week: How many calls and emails could your customer have resolved with the right documents, and how many really needed an expert? That number is the basis for everything else.
Three levers that work without new hires
Prevent requests.
Your customer finds documents, the parts list and the right part on their own, where they work. Don't expect miracles: only 14% of customer service issues are fully resolved in self-service, and 36% of very simple ones (Gartner, survey of 5,728 customers, 2024). In Transaction-Network's example for a customer portal, about 26 of 100 requests reach the portal in the first years, and about 7 are fully resolved there, between 3 and 11 depending on the company.
Make every case complete.
Whatever your customer can't solve reaches you with the machine, the serial number and the issue. Your team skips the follow-up questions and can start right away. This lever works on every single case, including the hard ones that land with you anyway.
Make knowledge available to everyone.
Answers from manuals and service documents are available to the whole team, including the new hire. In a study of 5,172 customer support agents at one company, agents with an AI assistant resolved 15% more cases per hour, and less experienced agents 30% more (Brynjolfsson, Li and Raymond, Quarterly Journal of Economics, 2025). Whether that carries over to servicing machines has not been measured.
No lever replaces experts. In a survey by KVD, the German customer service association, 77% rejected the idea that digital assistants would make up for the skills shortage in service by 2025 (KVD TrendRadar 2022, 154 responses). The goal is more modest and more realistic: your experts spend their time on the cases that need them.
What has to be in place
Your installed base.
You know which machine with which serial number is at which customer. Without that, your customer can't find anything about their machine, and your team keeps asking.
A parts list per machine.
Your customer needs the parts list of the machine on their floor, with successors for discontinued parts. If it is incomplete, they don't look for the part in the portal, they call. Often that is the real work, more than the software.
A place your customer actually uses.
Fewer than 30% of purchasing decision-makers use their suppliers' digital offerings, partly because the value isn't clear to them (McKinsey, 250 respondents in North America and Europe, 2025). Your customer runs machines from many manufacturers. A portal they have to open separately for each manufacturer gets opened only when they look for you specifically.
A measurement.
Count each month how many orders and requests arrive without a phone call and how many follow-up questions your team still has to ask. That is the only way to see whether service grows with your installed base while the team stays the same size.
How Transaction-Network does it
Transaction-Network is vendor-neutral software for servicing machines. Your customer manages all their machines there in one account, including other manufacturers', and finds your parts, your documents and a direct line to your team at your machines.
Ordering from the machine record.
In the Asset Management module, every machine has its record with parts list and documents. Through E-Commerce, your customer orders from you right there. You load spare parts kits per serial number by CSV or interface, and a part's successors by interface.
Service Requests instead of calls.
Your customer contacts you through Service Requests, including requests for a quote or for parts.
The Service Brain.
It answers questions from the documents you upload. When you upload a document, you decide whether the Service Brain may use it.
See who is on board.
Your customer list shows which customers are active and which are not, and you can filter by it.
Questions
Frequently asked questions
How many service requests can self-service absorb?
Fewer than often promised. Across industries, 14% of issues are fully resolved in self-service, and 36% of very simple ones (Gartner, 2024). In Transaction-Network's example for a customer portal it is about 7 of 100 requests, between 3 and 11 depending on the company.
Does AI replace skilled service staff?
Not as things stand. What has been shown: an AI assistant in one company's customer support raised resolved cases per hour by 15%, and by 30% for less experienced agents (Quarterly Journal of Economics, 2025). The hard cases stay with your experts, and they have more time for them.
Where do I start?
With your installed base and the parts list per serial number. Then count for a week which requests wouldn't have needed an expert. That is where you start.
Does my customer have to roll out more software?
Not with Transaction-Network. Your customer works in the browser in one account where they manage all their machines, and their maintenance system can stay.
Read next
- Knowledge
Does a customer portal pay off?
No reliable percentage, but a way to calculate it: how little has to change, by revenue.
- Knowledge
After-sales software for machine builders: types and selection criteria
The selection: the types available, eight criteria for your evaluation and what to ask every vendor.
- CEO Straight Talk ·
Why service is now the bottleneck in machine building
Machines are getting more capable, plants more complex. Service is not keeping up. That is rarely about technology and almost never about people.
Demo
See your machines in one account.
In a demo we show you what your machine park looks like in one account: every manufacturer, every machine with its record, maintenance and the Service Brain.