Key takeaways
The word that sticks
There are words that stay with you after a conversation. In my conversation with Risto Pfalz it was: swivel-chair interface.
It refers to the person who sits in front of two systems, turns from one to the other, and transfers information. Copy. Enter. Check. Forward.
Actually remarkable that in 2026 we still need a word for this at all. But a surprisingly large number of people in operations probably know exactly this kind of work.
A new tool for every problem
Risto has been working with logistics, e-commerce, and operational processes for decades. Today he advises companies primarily on optimising existing operations. In doing so, he observes the same pattern time and again.
A company has a problem. So a tool is introduced. Then there is another problem. So the next tool arrives. The first problem may be better solved — but at the same time a new interface has emerged. An additional process. Another data source. Perhaps even new manual work.
At some point the operation consists of a landscape of systems, processes, and workaroundsthat has grown over the years.
"You solve one problem with a tool — and in doing so, frequently create at least one new one."
The problem may then no longer be a lack of technology. But rather that no one steps back and asks: does the operation we have built actually still fit our business?
Operation first. Software second.
This idea runs throughout Risto's work. When he analyses a warehouse, he does not start by trying to introduce as much new technology as possible. He starts with the operational task.
How do goods move? How do orders move? What different types of orders are there? What processes do we need for them? Where does effort actually arise?
Only then does the question arise of what technology is necessary for this. That sounds obvious. In reality, however, the opposite frequently happens.
"The WMS does not replace the need to have a conceptual idea of how you want to run your warehouse."
Software can support processes. It can automate. It can enforce certain workflows. But it does not replace the understanding of how an operation is actually supposed to work.
Perhaps communication is sometimes a warning signal
Another statement from Risto I found particularly noteworthy. In companies you often hear: "We need to communicate more with each other." Risto's reaction to this is almost the opposite.
"You only need to talk when something isn't working. And the more you talk, the bigger your problem."
Of course, the point is not to abolish collaboration. The point is a different one. If employees constantly have to speak to each other to find out what has happened or what needs to happen next, then communication may be compensating for a poorly organised process.
A well-organised operation should, under normal circumstances, run almost unremarkably. The vast majority of the business should simply "purr along". People's attention then belongs to the few percent where something genuinely isn't going to plan.
People should resolve exceptions — not connect systems to each other
In order management and logistics, people still spend an extraordinary amount of time moving information. An email arrives. A booking changes. A supplier sends an update. A deadline shifts.
Someone checks one system. Then another. Then Excel gets updated. Then someone receives an email. And somewhere in between, there sits our: swivel-chair interface.
Yet the real value of these people lies somewhere entirely different. Not in transferring information from A to B. But in recognising when something is wrong. Understanding an exception. Setting priorities. Making a decision.
Risto too sees exactly this as the more meaningful role of the human: routine data entry can increasingly be automated. We need human intelligence where data shows something unusual — and someone needs to recognise: We need to look more closely here.
Perhaps we don't need more software
After my conversation with Risto, the question that stayed with me was therefore less about which new tool operations still need. But rather a different one: How much of our current operational work actually exists only because our processes and systems are not well enough connected to each other?
Perhaps the next generation of good operations does not begin with yet another dashboard. Not with yet another tool. And not automatically with AI either.
But with truly understanding the operation first. What is supposed to happen? What information is needed for that? What can run automatically in the background? And when is a human actually needed?
If we answer these questions correctly, we might perhaps one day be able to retire one of the most evocative German words in the world of operations: the swivel-chair interface.