“Then the organization is also more mature today than before and thinks more digitally,” says Verlage. “And with the help of this, I try to set the stage, let it out in the business, and have control only where you need to have control.”
Creating digital twins
To get the data in, Munters creates a digital twin for each machine it sells, and it also makes sure it’s clear it owns the data that’s generated.
From the moment an order for a machine comes in, the digital twin is born where all specifications are included. Then it follows through the production of its physical twin and knows exactly what fan and motor are inserted, and the serial number. It also has the information about where it’ll be delivered and when it’ll arrive.
“We get use of the digital twin because it keeps track of everything happening on time, and can warn if there are delays,” he says. “It’s very important for many customers.”
The digital twin also follows when the physical twin is installed — knows who’s doing it, notices deviations, keeps an eye on software, and then sends data from the physical part into the digital one.
“We only need to apply intelligence to the twin and not have it in the machine,” he adds. “It simplifies installation and maintenance, and we can handle everything comfortably at home from our servers.”
Safe from hacking
There’s also a security aspect to it all. For machines with sensitive information, it’s possible to ensure there’s only information going out to the digital twin, but it’s not possible to send information back to the machine to avoid it being hacked.
“If it were possible to shut down the cooling in a data center because we’re hacked, that would be a bad day at work for me,” he says. “So in some places, we might have a physical barrier where we need to go outside to apply intelligence to a machine inside the firewalls.”
Data coming in from all the digital twins can bring quick wins such as being able to save energy by setting up the machines as precisely as possible — an area where there’s still a lot to do, according to Verlage, who explains that today, it’s good to set machines with added difficulty to be on the safe side. But when you have accurate data, it’s also possible to set them exactly, so no extra energy is wasted.
But perhaps most importantly, if all the data that comes in can eventually be analyzed with AI, Verlage expects this can lead to many new insights and business opportunities.
Three guiding stars
Verlage has three guiding stars for how he should work with AI. The first is with agile to develop, implement, and go into operation quickly. The second is a methodology that comes from Google — pretotyping.
“Here, the ‘pre’ stands for pretend; you don’t have to have a prototype,” he says. “It’s enough to pretend you have one. Talk to the customer about something you don’t have but can do. If you’re going to test a new cell phone, make it in the shape of a wooden block and let people feel it in their hand and see if it initiates a spark.”
And the third star is what’s called “radical simplicity,” to make everything as simple as possible. To launch small, very simple applications and then improve them together with the customer.
Verlage says Munters recently built a customized customer portal where orders could be seen and tracked. And in one instance, on day two, the customer discovered an order was on its way to the wrong building and it was stopped and fixed.
“It was a wow experience for the customer, and shows that even simple things can quickly come in handy,” he says.
Leave the old
This approach to get things done and out quickly is very important to Verlage, and he thinks it’s often been strangely difficult for the IT business to cope with. Instead, large systems are often implemented for a long time, yet neglected.
“We can’t allow such waste and introduce expensive systems that aren’t used,” he says. “If we do, we’re doing something wrong. We have to leave the old world and digitize in a different way into the future. I’m almost obsessed with getting things into production.”
