AI in SMEs – “We have to seize the opportunities.”

Mr Würtemberger, tell us a little bit about Marabu.

We are best known for our artist supplies, but actually, that’s become the smallest division. These days, our focus is more on industrial applications—we decorate toys, have a foot in the smartphone industry, make keyboard labels and much more. We’re involved in everything from white goods to aircraft decoration.

How would you define your role within Marabu?

It’s my job to drive innovation, and to do so, I talk to a lot of people every day looking for the answer to the question: “How can we help you?”

Implementation is also part of that ...

Of course, but that is generally handled by my team. I focus more on strategy and mindset. We want to experiment, test out ideas and try something new.

Marabu does most of its experimenting in the lab, is that right?

That’s right. We work very closely with chemists, who develop paints for all kinds of use cases.

Paint a picture of what it’s like in the lab.

Well, there are people in lab coats, test tubes and lots of raw materials, water and pigments, and the people are mixing and innovating.

They write everything down in a little black book.

Actually, they don’t any more. We’ve replaced the books and Excel spreadsheets with interactive, AI-driven software so we can use the knowledge gained throughout the company and accelerate growth.

Like a digital paint encyclopaedia?

Exactly. We’ve christened the software ‘Albert’ in honour of our founder. While Albert makes suggestions to our chemists, it is not a replacement because we understand it’s essential to have the gut feeling and expertise of a real person, too. The AI is actually always learning from them.

Do you need hours, days or weeks to develop something new?

That depends. It takes anywhere between 10 and 100 attempts before we have success, and that can be over the course of several months. You’d think paints and inks were quite simple, but our customers have so many requirements—from scratch resistant and water-repellent to resistant to heat and so much more. And these aren’t things we can develop overnight.

Is AI reducing the number of attempts?

Absolutely. We’re only just starting out on our journey with the technology, but we’re already seeing the impact. When we can reduce the number of attempts from 100 to 80, we’ll need fewer raw materials and we’ll be able to deliver the finished product faster—two benefits that we shouldn’t underestimate.

And how did you get the team on board with AI?

We didn’t have to. The drive came from them. We made sure our colleagues in the development department were included from the off in a project that became so much more than just about the software.

Could you elaborate on that for us?

Ultimately, it was a huge restructuring project. Workplaces have changed. Warehousing has changed. The labs have changed. Among other things, that means the equipment we use and how we work and AI has helped us digitalise across the board.

Let’s talk about the workstations.

Where once there were PCs and keyboards, now there are projectors projecting interactive graphics onto the table. Of course, the chemists use raw materials and solutions, which means we were having to replace their keyboards and monitors fairly frequently. Now, however, we have installed touchscreens with covers that can be removed and replaced as needed to maintain cleanliness. 

But that’s not all, is it?

No, we have also introduced VR smart glasses. When a colleague in Tamm develops a new formula, they can collaborate in real-time with colleagues in China.

How enthusiastic are your team about these new possibilities?

Very. We’ve even managed to win over those who were a little sceptical to start out with as they’ve grown to appreciate the extra efficiency and opportunities the tech affords.

Are you and your team able to pinpoint the benefits of introducing AI?

We couldn’t give precise details—we have seen plenty of improvements—but I’m sure we are being held back by constantly calculating the ROI. Our aim is to be pioneers in AI use by 2029. It’s the only way we will be able to produce to the standard we expect in Germany.

What’s the urgency?

SMEs need to seize the opportunity to introduce quick decision making as well as a start-up mentality to remain competitive.

So your advice to other businesses is to jump right in?

I’d say it’s crucial that they do. Try it out a bit first without worrying how you’ll roll it out on a grand scale. The world of AI is changing at a rapid rate of knots, and so it doesn’t make sense to “wait and see”. It is important to realise that not everything will work as expected, but you have to learn from it and apply what you gleaned to the next use case. We have to be brave, but also willing to accept the need to reassess what we are doing during a project. Think of a child—it starts off at nursery, then goes to school before eventually heading off into the wide world. That’s exactly how you should view AI. You shouldn’t expect to get a fully-fledged graduate.

Children have teachers. How big a role does training the AI and employees play?

Both are extremely important. AI needs constant training and I also have to keep adapting the models to find the best solution, but for me, training the employees is considerably more important. They have to understand the fundamentals behind how AI works, learn the prompts and how to interact with it. But that’s not all as they really need to know that not everything the AI spits out is 100% accurate. We’ve put together an e-learning course and have also developed our own programs, because without these, we’d never progress.