“Thinking based on value, not data”

The importance of IoT is growing, especially in manufacturing. Manufacturing companies are increasingly equipping their machines with connected sensors. These in turn provide a growing volume of data. “But data without insight is worthless,” says Sander Huysmans, Sales Director Enterprise & Growth at SAS. “Once you are caught in a data swamp, it is difficult to get out of it. Equally, data without action is also worthless.” That is confirmed by consultancy firm McKinsey. They showed that seven out of ten IoT projects that used AI delivered results that exceeded expectations. Without AI, that figure is significantly lower.

“IoT is becoming more and more established,” says Dries van der Kleij, Business Development Manager Industrial Analytics at Proximus. “The building blocks are available: sensors, cloud, AI. The real challenge now is to look for the real added value.” AI allows manufacturing companies to maintain their competitive edge over, for example, competitors from China. “AI then provides, among other things, the necessary solution to speed up processes and make them safer.”

Start with the end point

According to IDC research, the combination of IoT and AI should lead to a reduction in cost by ten percent. So why are not more companies making that bet? “They do, but too often they start from technology,” says Sander. “That’s exactly where things go wrong.” Eight out of ten of the algorithms never get past the pilot phase. For the remaining two, putting them into effective use takes at least seven months. “You should not look at the technology as the starting point, but instead use the desired outcome as the end point.”

Sander Huysmans, Sales Director Enterprise & Growth at SAS

“Clearly, the applicability of AI cannot possibly provide equal value everywhere,” says Dries. “A great deal depends on the conditions in the plant, on how the processes are organized, and so on.” This only makes it more important to start based on a certain strategy, with a clear goal in mind. “That requires a lighthouse case,” Sander adds. “A concrete project that you can use to get the whole company on board: preferably not too technically complex, but above all with a clearly visible added value.”

Such a route allows starting at one production line, for example, and then expanding further. Sander gives the example of a project with SAS software at a company in the chemical industry. “The combination of IoT and AI resulted in production volume increase of 20% at three production lines, with a 26% reduction of energy consumption. That shows how the company looked at the project from a concrete objective: more results at lower costs and in a more sustainable way.”

A wide range of applications

There is no shortage of possible applications. “Everything you can think of as a human being can be put into an algorithm,” says Dries, “from very simple models, for example around anomaly detection in a production environment, to deep learning and neural networks.” In practice, companies today mainly opt for pathways that, for example, improve the usability of machines or the quality of products. “There are also great cases to take from logistics. The combination and analysis of large data volumes there helps optimize logistics goods flows.”

Edge computing and 5G as game changers

In addition, technological evolution in itself also creates new opportunities. “Communication sometimes used to be the slowing factor in IoT,” says Sander. “Edge computing plays a part in removing that barrier.” That is because edge computing offers the opportunity for data analysis to be done locally. Yet here, in particular, we should see 5G as the real game changer. “With 5G, there is virtually no latency any longer,” Dries explains. “5G allows IoT applications with large data volumes to run mobile and without any latency, even in environments where a wired solution is not possible.”

Dries van der Kleij, Business Development Manager Industrial Analytics at Proximus

A strong partner ecosystem

It is of course no coincidence that SAS and Proximus are joining forces. A successful IoT project rests on two pillars: data and analytics. “When you start a project, you simultaneously have to include those two aspects,” says Dries. Proximus’ connectivity enables data to be brought together and then SAS’ applications extract the necessary insights.

Microsoft – a partner of both companies – fits in here as well, as a provider of, among other things, the required storage and computing power. Codit, one of the companies in the Proximus ecosystem, is in charge of integration. “That is how we take the complexity out of the process,” concludes Alex Lorette, Director of Enterprise Solutions & Platforms at Proximus. “For the customer, this leaves only the most efficient approach, supported by our partner ecosystem.”