Overcoming the AI pilot trap: The VDMA Congress 2025
The third VDMA Congress, one of the most important platforms for technological dialogue in the DACH region, took place in Vienna on May 8 and 9, 2025, under the motto "Digital Solutions for Mechanical and Plant Engineering." Netural was among the 300 participants.
With more than 3,500 member companies, the VDMA (German Mechanical and Plant Engineering Association) is one of the largest industrial associations in Europe. Their events attract people from industry, research, and digitalization. We were there, gained inspiring insights, and derived five key takeaways, which we share here.
1. Why companies fall into the pilot trap.
Many companies are experimenting with AI, but few are putting projects into practice. According to a study presented at the congress, while 91% of the companies surveyed are planning investments in generative AI, only 7% are already using these technologies productively.
The reasons for this are varied: Suitable data is often missing or scattered across different systems. Many companies remain stuck in the so-called 'pilot trap' – they've tried something, but without a clear strategy or sustainable plan, it remains an attempt. Furthermore, top management commitment, which is crucial for anchoring AI initiatives in the long term, is often lacking.
In addition, cultural and organizational factors are still massively underestimated when introducing AI. Successful digital transformation requires not only technology, but also change management, clear communication of goals, and employee involvement from the outset.
Netural@VMDA 2025
2. Digitizing knowledge to address demographic change
A recurring theme: Knowledge about complex machines and processes is often held by long-standing employees. In the coming years, the baby boomers will retire, threatening a massive loss of knowledge. When they leave the company, critical information is also lost. Therefore, many companies are investing in digital knowledge systems that provide structured access to information from emails, service tickets, and conversations.
Sandra Szech (Odego) and Stefan Hötzinger (Scheuch)
3. Reduce planning and sales cycles with Agentic AI: Scheuch shows how it's done
A highlight of the conference was the presentation by Scheuch, which demonstrated how to develop marketable tools from pilot projects. In process engineering for gas purification plants, Scheuch faced the challenge of planning highly complex individual systems in the shortest possible time.
The answer: A two-stage software system consisting of a modular system for plant configuration – and the Process Designer, which we developed at Netural. This application, specially developed for Scheuch, automates the process engineering design of the plants and thus forms the technical heart of digital engineering.
Data quality, interfaces, and Agentic AI – all of this is already being used productively here. A practical application that shows: AI projects start with good data ("Data is the new soil") and a real need.
4. There are already many concrete use cases
Several presentations demonstrated how digital twins and AI are transforming the entire industrial value chain: from sales, planning, and design to after-sales and service. This creates end-to-end digital processes that already simulate, predictively control, and optimize real-world systems.
A concrete use case from heavy industry demonstrated how sensors, acoustic data, computer vision, and explainability interact in an AI-supported control system. This presentation was presented by our customer Primetals, which impressively demonstrated how AI can be used not only for process optimization, but also for knowledge retention and decision support. The examples ranged from acoustic diagnostics and visual feedback systems to explainable AI for complex trend analyses.
5. Data is the new soil
One of the most frequently mentioned topics at the conference was the importance of data and its quality. Several speakers emphasized that data-driven AI applications can only function successfully if the underlying data is valid, structured, and available.
Kurt Herzog from Primetals summed it up: "More than 80% of the effort in data-driven projects is spent on data preparation." Stefan Hötzinger from Scheuch also emphasized, with the motto "Data is the new soil," that good data is the foundation for every successful AI application – only from this can sustainable added value be generated.
Our Conclusion
The VDMA Congress impressively demonstrated: Digital transformation is no longer a testing ground. Those who want to be successful today must overcome pilot projects, secure knowledge, use AI effectively, and scale digitally.
At Netural, we have developed our "AI Challenging Team" that addresses precisely this. Our team works with our clients to analyze the real potential of AI, identify concrete benefits, and support the implementation. Want to learn more? Then contact us or get started directly with the AI Challenging Team.
In addition, we have published a white paper: "How AI Really Works" – a compact guide for anyone seeking working solutions, not castles in the air.
