At the Symposium this year, we were represented with a large, highly visible booth and a strong team to strengthen our position as a leading Digital Thread Expert. One of the dominant topics in customer conversations at the booth was not so much the use of AI itself, but rather the question of how we can support companies in unlocking their data treasures and connecting them in such a way that they can actually be used by AI. What became clear at the Symposium during the keynote at the end of the first day was that data must be structured and interconnected to be enriched with context and made queryable.
“Agentic AI and the Rise of the Software Defined Industry” was the ambitious motto of the event, which the main sponsors brought to life in their keynotes. Patrick Müller and Frank Patz-Brockmann (CONTACT Software) emphasized that companies would not win through AI itself, but through the way they use AI – and for that, the context provided by their IT systems is essential. PLM, they argued, is the foundation for AI. Wolfgang Puntigam (AVL) likewise stated that AI agents are always only as good as the data they have access to. Used consistently, 80 percent of activities in research and development could in future be supported by AI – if not even automated.
The problem with this, however, is skill erosion, which Patz-Brockmann drew attention to. If AI takes over the tasks of the junior engineer in the future, we will soon no longer have engineers with the necessary experience to control the AI.
As people, companies, and society, we therefore have to learn to handle AI responsibly, as Martin Eigner (Eigner Engineering Consult) and Helena Gutiérrez (Share PLM) made clear in their presentation on Humans as the Real Differentiator. No easy task, given the enormous speed at which AI is evolving: ChatGPT reached a query volume within a short period of time that it took Google around ten years to achieve.
Three Presentations and a Workshop on the Topic of Openness
Replicating the complexity of human communication is, despite ever better language models, a challenge that cannot (yet) be easily mastered with AI. And communication only works if people can trust each other, as Knut Stang and Katrin Fritz (Volkswagen) explained in their presentation on group identity and language. People do not communicate in a uniform language, but in different dialects and in professional- or industry-specific expert languages. Training ChatGPT & Co. to correctly interpret this will not be easy.
Knut Stang was not our only speaker from PROSTEP at this year’s Symposium – in total, we were represented with three interesting presentations. Marvin Höhn from BHC, together with David Leuckhardt (Mercedes), explained how the faster cycles of software development can be aligned with the development of the corresponding electronic hardware in such a way that an unnecessarily large number of hardware components do not have to be replaced during testing. The basis for this is the linking of data from the E/E PDM system with the software releases in the MB.OS portal to form a Digital Thread. This not only saves several hundred thousand euros in costs for unnecessary modifications but also enables much more intensive testing.
In another presentation, Martin Holland, together with Jörg Hoffmann (BOS), presented a solution developed as part of the Decide4ECO project that bridges the last mile between data ecosystems and companies’ backend systems – in order to share sustainability information. The demonstrator of the solution is based on our OpenPDM integration technology, which we have extended with an EDC Connector for connecting any data ecosystem. In this context, OpenPDM also takes over control of the policies for access to the data and its use. Furthermore, as part of the project, we have developed an AI agent that consolidates sustainability information from various sources into a Digital Product Passport.
A prerequisite for building such cross-system Digital Threads is open IT systems and support for standards. The use of data ecosystems such as Catena-X or Manufacturing-X and the deployment of AI-supported tools also only works with open systems. Unfortunately, openness has tended to decrease rather than increase with the trend toward cloudification and the transition to rental models. This was one of the key insights from the 90-minute workshop on the topic of the Code of PLM Openness (CPO), which was coordinated by PROSTEP. Conducted in the style of a World Café, the significance of openness for the industry and the main obstacles were lively discussed at three stations. The discussion is to be continued at the next Symposium.
The next prostep ivip Symposium will take place on April 14 and 15, 2027 in Düsseldorf.
Best to mark the date in your calendar already.