In the age of smart manufacturing, we would expect manufacturing processes to be more holistic and seamless.
Unfortunately, in the manufacturing industry, siloes often appear across departments, teams or systems, adding another dimension to the challenges that manufacturers face in the fast-moving digital economy.
These siloes may be functional gaps between departments like production and quality control, inter-plant silos between different facilities, or data silos where information is stored in isolated systems.
Siloes are stumbling blocks to effective communication and collaboration, leading to inefficiencies resulting from missing information, slow response times, missed opportunities for improvement, and redundant work.
At the recent Siemens Realize LIVE 2025 in Detroit, DigiconAsia caught up with Tobias Lange, Senior Vice President, Manufacturing Operations Management, Digital Manufacturing, Siemens Digital Industries Software (SDIS), to discuss the challenges in the manufacturing industry in Asia Pacific, and the role of AI and data analytics in creating the smart factory of the future.
What are some key challenges in digitalizing traditional manufacturing processes?
Lange: The manufacturing space is very complex. Besides having to deal with brown field issues such as sustainability, especially when it comes to energy consumption, and other global trends such as labor shortage, aging populations (this is a real challenge in Korea, for example) and geopolitics impacting geographical manufacturing strategies, manufacturers have also to consider customer demands in their digitalization journey.
Customers today want products right now, but they also want them personalized. They want to have an experience that is tailored to their needs, and that has implications across all industries.
How to get all these done is already a challenge, and what’s more from a regulatory perspective – what kind of processes you have to follow in order to document what kind of materials you’re using, what kind of carbon footprint, the materials that you’re using, the traceability and so on – the challenges have taken on a new dimension.
While design, manufacturing and operations should work together holistically and seamlessly, without siloes, we find that sometimes quality is siloed – when it should be glove-in-hand with operations and design. Such siloes are also a key challenge to digitalization.
It’s good that one of the trends we see in a lot of very successful customers is that these silos are getting broken down, and that the teams are working more and more hand in hand.
How does SDIS envision the ‘smart factory’ of the future?
I think the important thing is that factories will look very different depending on the use cases, depending on the needs; thus, the factory also has to be adaptive. One of the key elements that we will see in the future – and actually we see some of it right now already – is the need to be adaptive to process changes right down to even the kind of products that you’re producing.
You have to be adaptive, so that when you want to produce one product in one factory one day and then you want to do it differently the next day, you want to be flexible around, for example, supply chain scheduling and shortages. You want to be flexible in the way that you adapt to processes.
This has to be in an environment where there’s much more collaboration between machines and humans. There are going to be industries and factories that will be completely lights off – only machines without humans – but there will also be a lot of places where you still need humans.
Meanwhile, these humans need stronger assistance and help in the workforce from machines and software, which in turn need to have a have proper guidance from humans…
It’s all about the environment, about the setup of those factories, what you can envision for the workforce, and the collaboration and support that that we can provide, from a software perspective, to meet the needs of the companies that actually are producing things.
One key thing is the connection between software and humans in any particular environment.
In what ways do AI and data analytics help with quicker turnaround times, reduced costs, and improved decision-making in manufacturing?
Lange: A lot of times we see, as an industry, people sprinkling AI like a magic dust on top of everything and think this is a solution.
Even though there are lots of trends changing the industry and the future of factories, the core value proposition and use cases that manufacturers want are: quicker introduction of new products, increase in yield and quality, and lower cost of production.
So, it’s all about finding and identifying the right use cases and really solving those use cases.
For instance, we see that you may have a lot of documents like standard operating procedures, recipes and so on, and distilling all that down into a standardized bill of process takes a lot of time. We’ve built solutions that can help decrease the time required from like six months, depending on industries, to maybe two weeks.
We have customers that build rockets that go to space, and they have lots of complex requirements that they have to solve. They want to understand, if they insert a certain part into a certain place in the rocket, what is the procedure that they need to follow, or if something goes wrong, they need advice as to how to repair it. Again, there are lots of documents they would have to search through. AI can help speed these things up.
Another area is in scrap reduction. It’s about lowering maintenance cost, understanding how you prevent machines from breaking down, and having predictive elements to prevent failure.
Ergonomic studies can be very costly, but AI can facilitate and reduce the cost of such studies and the analysis. For instance, instead of somebody actually doing the overall modeling for ergonomics, you can just take a short video, send it to our system and analyze that, and understand that there are certain issues with the way that that a user is lifting something. So maybe we need to position the inventory differently to help workers not have back injuries after a year of working!
These kinds of things are important. And we have to look at them from a use-case perspective. Obviously, these examples are more from the shopfloor, but you can take that to the next level: what if you are a shift manager, and you want to analyze certain flows in the factory and get better throughput, when you know that certain batches in production have certain issues, and how to resolve that – maybe from an engineering perspective, all the way back from the operational workflow to the design of products.
Also, AI can help analyze factory networks, helping the Head of Production look where certain factories do certain things differently, and because of that, they have a higher yield.
One of the core elements we see in the manufacturing world is around material usage. It has become important to see that we have end-to-end traceability, that restocking happens when it’s necessary, and to have, for example, the right labels displaying the right content (especially where food manufacturing is concerned), and seeing that the entire flow is getting traced.
I think there are lots of possibilities to utilize AI to analyze this kind of flows and be able to take the capabilities of AI to analyze and optimize the product design process. With the relevant data, for example, you might want to make changes to reduce the amount of materials used in manufacturing a product.