AI Design Week Provides Logiqs with New Insights and Practical Tools

As an industrial equipment supplier, it is crucial for Logiqs to keep up with the times and stay abreast of the latest trends and developments. The possibilities offered by artificial intelligence are of particular interest to the company, which is why it wants to get its developers excited about working with it. It is also interesting for them to develop an ecosystem in this field with partners and employees, both for their traditional systems (for example, for logistics and greenhouse farming) and for future projects (vertical farming). In their search for a party to help them with this, Logiqs quickly came across Info Support. The Design Week in particular appealed to them.

What makes this Design Week so special is that Info Support’s experts spend one full week working with the client’s developers, in this case Logiqs, on a prearranged assignment. “During this week, we devote our full attention to the assignment. A huge advantage,” says Téa Stojanovic, IT consultant at Info Support. “When you start working on a new project, new elements or issues always pop up. If you include those in your process, you end up straying from your goal. By focusing completely on that one hypothesis and not having time for ‘strays,’ we stay right on target. We do write down new findings so we can address them another time.”

“The limited duration of one week actually made it a no-brainer for us to participate in Design Week,” said Marijn, who didn’t participate in the week himself but guided the process from Logiqs. “Today, with software, you have to be able to put something down very quickly and learn from it at a rapid pace. If something like this takes several weeks, you may find out at some point that you should have done something differently in your first week. By doing everything in one week, you are able to be more flexible with this. In addition, it’s easier for us to schedule some employees for five days completely off so they can focus on this properly, without the distraction of other work.”

Before Design Week starts, Info Support’s experts and the client’s developers meet for a half-day session to brainstorm ideas and hypotheses that will be addressed during the week. Of course, the client’s wishes and possibilities are central to this. Ultimately, one assignment is chosen together. In the case of Logiqs, it was about predicting when a particular failure would occur so that they could anticipate it.

After the developers first attended another three-day AI training course, they began their AI assignment in collaboration with Info Support. This involved the Info Support people setting up and training an AI model, while the developers enriched and cleaned up the necessary data. “Unfortunately, on Thursday, the fourth day in other words, it turned out that we were not going to be able to predict the failures,” Melvin Ducaneaux, senior software engineer at Logiqs and an AI Design Week participant, told me. “First of all, disappointment prevails because after all, you are all working hard on a solution and then it is annoying when it turns out that what you want is not possible. But soon the realization dawned that this intensive week had actually been very successful because we had gained so many new insights and learned so much in the field of AI.”

When it became apparent that it was not possible to predict failures based on current data, the Info Support and Logiqs team set out to find what was needed to make this model work. That was then put on paper so the developers know what they need to take next steps toward an even better product in the future.

In addition, it showed the need to involve AI experts in any project. A three-day course is useful and teaches the basics, but obviously does not provide the knowledge that experts have already gained during their education and careers. Therefore, during Design Week, it was nice that Logiqs developers could build on the expertise of their teammates from Info Support.

“To be honest, I’m kind of glad it didn’t work out to a concrete result,” Marijn said. “In fact, I would have become very suspicious if it had been one big success story, because obviously the challenge was not big enough. For us it was mainly about the learning process and based on these results we can very well build further. For example, we have now realized that we can do about three-quarters of an AI project ourselves. For the remaining 25 percent we enlist the help of relevant experts, which is why it is so important for us to build a good network in this area. So that we can always enlist the in help of expert partners.”