Jens Beck

Partner for Data Management & Innovation at Syntax


What are some ways you see artificial intelligence of things (AIoT) disrupting the manufacturing space? What are some emerging opportunities for the technology in 2023?

The concrete examples I see every day are mostly predictive cases. There is predictive maintenance, which is when the machine tells you it needs maintenance in advance to ensure you have your service technicians and spare parts in place. Then there is predictive quality, which allows you to predict the quality of the product — using sensory data — before it’s actually produced, reducing your scrap rate and rework efforts. 

Both directly impact your overall equipment efficiency and, therefore, your operational results. But while these two are quite common, visual inspection still has ways to go. Companies show interest, but implementations are not as frequent as with the other two yet. Visual inspection is basically quality inspection via pictures and AI models, so the camera is your sensor. The business case here is tremendous as technicians can drastically reduce staff workload and continuously inspect all parts. Now, of course, an area that is also taking up speed is collaborative robots, which are robots with sensors around them and some AI components to ensure human-robot collaboration.

One technology I really expect to take off in 2023 is natural language processing in the form of chatbots or assistants. Imagine your frontline worker or maintenance staff only needs to talk to their MES and ERP system and not open a complicated screen or application.

I really think that in 2023 we will see a lot of impact happening where AIoT is used to improve the human-machine interface.

With the rise of Industry 4.0 and connected devices creating more data than manufacturers can leverage, how can AIoT make sense of this business problem, which seems to be one of the barriers to the adoption of IoT?

Operational technologies used to have loads of data long before they had the toolsets to leverage that volume, but in time, the devices became chattier, which led to a wide range of business benefits.

So where is the struggle? In my opinion, the struggle has been the paradigm shift in data management and IT. The data sources of today are not fully structured, well described, nor relate 1:1 with ERP models. The snapshotting of data multiple times a day is not efficient anymore; you need real-time analytics, and last but not least, aggregating data has not been helpful in recognizing patterns, so efficient methods are needed to help store the data.

Research found that for the first time in five years, the number of cyberattacks in the manufacturing sector exceeded that of insurance and financial services. In being vulnerable to multiple system failures, the manufacturing space has been made a prime target for malicious actors. How can AIoT strengthen factory security?

OT networks are normally isolated from the outside world, as they should be. But I would identify AIoT as more of a risk to OT security if not properly applied, i.e., encryption, zero-trust communication, and so on. AI can also be used to increase OT and IT security. If we want to talk about physical factory security, then yes, AIoT has a range of benefits. If unauthorized people are strolling around, it can use facial recognition to raise an alarm. It can also protect workstations and terminals in this way.

The gear industry is heavily reliant on the automotive space. With the population’s energy preferences shifting, how do you see AIoT aiding in this shift to more eco-friendly options such as EVs and wind turbines? What do gear manufacturers need to know?

Well, with the shift toward EVs, wind turbines, and hydrogen, AIoT is even more important to accurately measure the state of your components, calculate on the fly the correct usage, and ensure the right mix of components to generate enough hydrogen.

Today, most energy options run on a service model, which is what the customer expects. They want to carry no risk but have 100 percent reliability that exactly aligns with their preferences. With this in mind, data is the fuel by which manufacturers can cater to their customers with deep personalization.

If you prefer a fast car, then the gears can be tailored to your specific car preference; the same if you want a slow car or a quiet ride. This is all enabled through data. So, the first thing gear manufacturers should consider is how they can collect and leverage data to adapt their products and identify potential areas for optimization. They need to close the gap between engineering and manufacturing both for business relevancy and to meet the ever-evolving needs of consumers. 

How can AIoT assist the gear industry in meeting the shifting demands of consumers from mass-scale production to more personalized, customizable products? Does AIoT bear any relevance to the gear industry here?

AIoT bears relevance in many areas, including the shift from mass-scale production to make-to-order products with high-variant configuration. I see a couple of consequences with this change. Certainly, testing and engineering will have to change to meet higher cycle times.

There will also be natural limits, so manufacturers must rethink testing and simulations. Sensors, AI models, predictions and probabilities, and therefore, AIoT, will play a major role there.

The same is true for production. Anything that can be done to guide frontline workers, optimize processes, increase uptime, and reduce scrap or rework is a plus. I know this to be true in the gear industry and with my own background in automotive. I know that gear manufacturers are up against the same challenges.