"The potential is unbelievable!"

Korbinian, everyone is talking about artificial intelligence – where do you see the greatest potential for companies?

It is assumed that about 15 to 20 percent of sales are lost in production due to quality problems. That's really massive. A company that makes 100 million euros in sales and loses 15 to 20 million of them due to material waste, rework, etc.: a huge topic. The other area is productivity: Managing increasing complexity, for example an increasing number of product variants, when do I produce what, which material has to be where at what time – there is also great potential here. In addition, there are other areas such as sales intelligence: At what prices do I buy, at what do I sell? However, the main issues remain quality and productivity.

 

Kineo works with GEALAN in some of these areas.

Yes, in production, quality and logistics. For example, in extrusion, our AI predicts how productivity will behave in the next 48 hours and how productivity would behave if I do a cleaning now, i.e. shut down the machine, clean for eight hours and then continue with higher productivity. The AI provides the cut-off point to decide: clean now or, better yet, move it – and live for all extrusion lines. A prediction that has not been possible before. In logistics, it's all about profile recognition via app – GEALAN has a lot of geometries, a lot of variants and AI makes the picking process easier. In the area of quality, it is about software that can detect and sort out deviations on the profiles much better.

 

Will the AI tools increase GEALAN's performance?

Yes, very strongly, in fact. You have to understand the complexity at GEALAN. Tens of thousands of profile variants, and the trend is rising, because customers want more and more choice. GEALAN has around 40 extrusion lines in Tanna. This simply exceeds any planning that a human can do. This is exactly the right point to start with AI.

 

Would you advise SMEs as a whole: where complexity and AI?

There is enormous potential in SMEs to build up data to increase productivity, but this is hardly being tackled yet. Data is only collected to support very central processes, for example to know which customer wants which product and when. What is not collected is: When and for how long is which product where? What quality do we measure where? These are questions that are impossible to answer on the basis of classic ERP data systems. So there is a lot of data that you first have to build up in order to develop AI solutions at all. If a company suddenly decides in five years that it wants to roll out AI on a large scale, then that is not possible. AI thrives on historical data, it needs data from four or five years. So I have to start building up this data base now so that I can draw from it in a few years – that's the big homework.

 

How did GEALAN complete this homework?

Very good! AI is not a means of production that I buy, write off with an amortization over a few years and then see how it pays off, but a cultural, transformative process. We want to take people with us and increase quality and productivity holistically. You can't impose that overnight. Many companies, especially in traditional industries such as construction, are still waiting and seeing what their competitors are doing. That's actually a problem, because at some point they'll just be behind and then try it with the crowbar. But then it is too late to start storing data. GEALAN is on the right track in this respect, has invested, is pooling its efforts. In my view, the management is approaching AI with a view to the future and I believe that this is exactly the right way.

 

Korbinian Kuusisto (35) is Managing Director and Co-Founder of Kineo. Born in Landshut, after graduating from high school, he studied economics in Hamburg, Helsinki and Berlin, then mathematics and was intensively involved in statistics. Like many statisticians, he then switched to AI and machine learning, worked as a developer for a start-up and founded Kineo with two colleagues in 2020. The Berlin-based software company develops tailor-made AI solutions for the manufacturing industry. The declared goal is to optimize productivity and quality in processes with AI. Kineo works for manufacturers of heating systems, shoes and soaps, for the plastics and wood industry, for car manufacturers, energy producers, banks and insurance companies. The company has around twenty employees: the developers are mathematicians, computer scientists and physicists; the average age is in his early 30s.

 

Why is Kineo a good partner for this? What distinguishes you from competitors?

First, that we offer Infrastructure-as-a-Service. Our partners don't have to form AI teams and build AI infrastructure themselves to experience the added value of AI. We provide you with the entire infrastructure to operate AI solutions cloud-based, including custom solutions. Secondly, that we provide the cultural component, i.e. that we really walk the path to AI together with the companies. We are on site, for example very often in Tanna and Oberkotzau. We take a close look at all processes and really want to understand them. We are honest in our assessment. For us, it's not about selling a solution come hell or high water, but about what added value we can achieve for a company. We call this "Genuine Consulting".

 

You talk about a cultural component – what do you mean by that?

AI needs a certain mindset, a certain open-mindedness. If I had tried to convince someone 30 years ago that we would soon no longer write letters, but e-mails, it would have been relatively difficult. You need an attitude to try things out and see where the journey is going. It is not our strategy to try to convince people of AI in general. But if you are willing to try it out and invest in innovation, we are the right partner for you. We have our tools and methods with which we then implement this well.

 

Good tools are one thing – what role does the human level play?

It is one of the most important things. In the end, everyone wants to work with nice people, whether they are external partners or their own colleagues. At GEALAN, we were very lucky because we really only met open-minded and simply very nice people with whom we work very well. It's also important to sit together in a meeting room and not just have remote calls – it just creates more consensus.

 

AI is an extremely dynamic topic. Is there such a thing as stability for Kineo?

Stability in the sense of leaning back certainly not. Elsewhere there are transformations in annual spans, we think in weeks.

"We started in 2020, two months before the stock market crash of Corona, but then grew consistently and healthily over the years. Of course, the global economy has put obstacles in our way – we are active in the manufacturer market – but AI is an area that is growing very strongly. Despite the crises, the discussions we have with companies are very positive. The willingness to invest in AI, in really good solutions, is there."

We are experiencing new technologies that Google, Microsoft or Chinese competitors are releasing that bring disruption to the market. Sticking with it is actually very challenging, even for a startup. When we started, for example, we couldn't develop some solutions because language processing with natural language processing wasn't that advanced yet. A few years later, the technology is suddenly years ahead. The same applies to image processing – there are problems today that we cannot yet solve, but which will soon be solvable. Everything happens incredibly fast. We keep track of things for the companies.

 

What does Kineo's ideal customer look like?

Medium-sized, i.e. with a certain volume, but not too strong corporate structures. The management thinks innovatively and is open to digitalization, data and AI. The people in the departments are willing to go along with this path. Products and processes are geared towards series production, high volumes, high material flow. They are perfect Kineo customers because we can really use AI with them. I could also answer the question more briefly: Our ideal customer is like GEALAN.

 

Will there still be areas of the company in the future that do not come into contact with AI?

Basically, AI is very good at doing things that very often happen in a similar way. AI will therefore play a greater role in series production than, for example, in toolmaking, where every tool is different – we call this quantity 1. But even there, AI will be extremely helpful in mathematical methods and in simulations. Artificial intelligence will sometimes be more visible, sometimes less visible, but it will be important everywhere.