Question: Everyone is talking about Digital Twin, Digital Thread and Artificial Intelligence. Is classic PLM outdated?

Schleich: No, in my view, "classic" PLM is not outdated. It is being expanded by the concepts you mentioned, which even today are already built very strongly on PLM systems as the data and process foundation, but add additional real-time, analytics and optimization capabilities. PLM remains the core data and process repository that enables all further technologies and will become even more strongly integrated in the future.

Question: The Digital Thread connects data in many domain-specific systems - does that not go beyond PLM conceptually?

Schleich: It is clear that PLM systems have to continue evolving and become more flexible so that they can connect more effectively. Nevertheless, PLM remains a building block without which, in my view, these concepts cannot function properly.

Question: Will PLM ever cover the entire product lifecycle, which is becoming increasingly long because of Software-Defined products?

Schleich: Particularly against the backdrop of the circular economy and the growing importance of product-service ecosystems, the classic, linear view of the product lifecycle is dissolving. Instead, product lifecycles are developing dynamically in parallel with the products; they are becoming a continuum, and PLM systems need to reflect this step by step.

Question: If you could reinvent PLM - what would you do differently, or what would vendors have to do differently?

Schleich: I see various starting points here. Many PLM systems have relatively monolithic roots, while today the platform concept plays a more important role. Modern PLM systems should be designed as Data Platform as a Service (DPaaS), providing all lifecycle data openly via standardized APIs - from CAD models and IoT time series through to AI models. And then, of course, there is the question of how data from product instances in the field can be integrated, how specialized AI agents can be incorporated and how we can anchor sustainability more firmly in PLM.

Question: Have we come a step closer to the idea of open, modular architectures in recent years? At the prostep ivip Symposium, it was claimed that cloudification has actually made platforms less open.

Schleich: That is certainly a valid argument, but I do see that quite a lot has happened in recent years. And I also believe that we will see innovations in the coming years, particularly from innovative start-ups, provided they are not immediately acquired by the major players again.

Question: Are today's PLM systems designed for data-driven, learning development processes, or are they more of a data graveyard?

Schleich: They were not originally designed for this, but they are a good foundation for data-driven and AI-supported development processes because they provide the opportunity to maintain the necessary data at a high level of quality. However, context information and therefore clean metadata also often play a major role in successful AI applications. In larger organizations, maintaining them is less a question of tools than of organizational structures and data governance.

Question: So structured data is better than a data lake from which AI can pick out whatever it needs?

Schleich: Yes. Even with large volumes of data, the data quality has to be right and has to be verified. Because they are connected to processes, PLM systems are well suited to ensuring this.

Question: Why are engineering processes still so poorly automated despite years of PLM use?

Schleich: Well, I have the feeling that many companies have set out on this journey and are now in the process of gradually automating their engineering processes. But that is more of a feeling, which I cannot substantiate. At the prostep ivip Symposium, however, I had the impression that a certain momentum has developed.

Question: What role will Artificial Intelligence play in automating engineering processes?

Schleich: We can see that the capabilities of AI methods are developing rapidly, and I am certain that in the future they will not only automate individual process steps but will partially automate complete development processes. At the same time, however, I believe that human intelligence and creativity will continue to be and remain highly important in the future. We need to build decision points into AI-automated processes at the right places, where humans make decisions and contribute their creativity.

Question: So you do not see AI independently generating new products in the future based on the available data?

Schleich: That is a vision that may perhaps be technically conceivable, but at the same time it raises many legal questions, such as who ultimately assumes responsibility. For that reason, in my view, it will remain important to involve human decision-making competence.

Question: Where are the biggest gaps in PLM processes with regard to digital continuity?

Schleich: It starts with the individual disciplines, for example with the integration of model-based systems engineering and the continuity of requirements management, and extends all the way through to production and operation, in other words towards ERP and MES systems and IoT platforms. In some cases, semantic gaps also still exist between domains, meaning that the same data and information are named differently. And then I also see gaps in the direction of AI, which is not integrated into the processes.

Question: Why is feeding lifecycle and operating data back into new product generations still so difficult?

Schleich: There are certainly still a few technical hurdles, but when I talk to companies, legal and organizational issues play the main role. Who is allowed to use the data, who is responsible for feeding it back, what quality is required, etc.? In addition, simply feeding back the data is not enough. Context information is often also needed.

Question: What are currently the most important research priorities in your department at TU Darmstadt?

Schleich: My team and I are researching how the product lifecycle can be represented in the digital space. We want to close information loops in order to improve product development as a whole and contribute to more sustainable value creation. In addition to automated information processing in tolerance management, our research focuses in particular on data-driven sustainability, the design of adaptable production environments and the automation of engineering workflows through the use of AI.

Question: Are there joint research activities, for example with the prostep ivip Association or with other organizations?

Schleich: Yes, my department is currently involved in the association's FAICE project, which focuses on integrating generative AI into collaborative engineering workflows.

Question: You have conducted extensive research into the Digital Twin. How many twins does a company need and in which IT systems do they live?

Schleich: This question cannot be answered in general terms. The number, scope and software- and system-side implementation of Digital Twins depend heavily on the use case and ultimately on the company's business model. After all, a digital twin is not created as an end in itself, but, for example, to support pay-per-use models or predictive maintenance offerings. Finding suitable use cases is a challenge, especially for smaller companies, but there are now providers that support them in implementing and also operating digital twins.

Question: Things have become quieter again around the topic of sustainability. Is PLM actually an effective lever for this?

Schleich: Topics such as sustainability and the circular economy have lost some attention due to the current geopolitical challenges. But I still attach great importance to them because, ultimately, this is not only about ecological aspects but also about cost issues and resilient supply chains. Given the need to collect and exchange sustainability-related data across the entire product lifecycle, PLM is certainly an important building block here, but it often has to be supplemented by additional systems, for example systems for material data management or IoT platforms in production.

Question: I will return to my opening question. Will your Product Life Cycle Management department still have the same name in ten years?

Schleich: There will certainly be some changes in the coming years, but I strongly assume that representing and controlling the product lifecycle in the digital space and closing data and information loops in engineering will still be of great importance in ten years' time. That is my bet, although I admit that I am biased.

Professor Schleich, thank you very much for this interesting interview. 

About the Person

Prof. Dr.-Ing. habil. Benjamin Schleich is Professor and Head of the Product Life Cycle Management (PLCM) department in the Department of Mechanical Engineering at the Technical University of Darmstadt. His research focuses on various aspects relating to the exchange of information and data along the product lifecycle, such as issues relating to data-driven sustainability, the automation of engineering processes, flexible digital manufacturing systems and information processing in tolerance management. Together with his team, he also works on Digital Twins, data mining and machine learning in product development, tolerance management (particularly in the context of Industry 4.0) and innovative methods and CAx process chains.