One of the most important prerequisites is openness

neue verpackung: Dr Eckertz, AI is currently a hot topic - but how specifically is it already changing the work of development engineers in mechanical engineering today? Can you give us an example?

Dr Daniel Eckertz: AI is changing the work of development engineers not so much through a fundamental upheaval, but through many small, very specific reliefs in everyday life. AI solutions are mostly targeted assistance systems for individual tasks. The greatest effects occur especially where a lot of knowledge is distributed, unstructured or has grown historically.

A typical example is dealing with technical documentation, specifications, etc., also from previous projects. AI can now provide developers with targeted information from large data sets or make suggestions for solutions, instead of having to search for a long time or start from scratch. In mechanical engineering, it is often about variants, adaptations and reuse of existing solutions. This is exactly where AI helps to make good technical decisions faster, based on experiences that have already been made and that one might not directly find again as a human, or at least only with a lot of time.

neue verpackung: In which phases of the engineering process - from conception through design to commissioning - do you expect the greatest changes due to AI?

Eckertz: I expect the greatest changes initially to continue in the early phases, i.e., in conception, design, and variant finding. This is where the central development artefacts such as requirements, concepts, and bills of materials are created - and mostly text-based, where LLMs in particular play to their great strengths. Today, however, these artefacts are usually not used consistently or only to a limited extent for the later phases.

Here, AI can help to use the information more consistently and to automate activities in a targeted manner at various points, for example in preparing production based on bills of materials or in planning tests based on requirements, etc.

At the same time, many companies hope for support from AI in the CAD area, because a large part of the development work takes place there. There are already initial approaches here, but the implementation is significantly more demanding, as it involves complex 3D geometries and technical relationships. However, I believe that we will see significant progress here in the coming years and that this can then change engineering much more than many of today's assistance solutions.

In later phases such as commissioning and service, AI will also gain importance, for example in the analysis of faults, deriving optimisations or supporting service technicians. The crucial point from my perspective is that the information generated in engineering remains usable throughout the entire lifecycle of the machine.

“AI is change management - at the speed of light”

Artificial intelligence is long more than a buzzword in the industrial environment. At the Automation NEXT Conference 2025, Dr. Axel Zein, CEO of WSCAD GmbH, showed from a manufacturer's perspective how engineering processes can not only be optimised but fundamentally changed through AI.

neue verpackung: You are researching innovation engineering at Fraunhofer IEM. Which AI applications do you consider to be market-ready, and which are still in the development stage?

Eckertz: For us, innovation engineering means bringing new technological approaches into application quickly and testing them practically at an early stage. This also deliberately involves applications that are not yet market-ready, to understand their potential but also their limitations. The goal is to prepare technical innovations so that they can be purposefully transferred into industrial processes. But there are, of course, already some AI applications that are market-ready.

On the one hand, these are general chatbots like ChatGPT. But also, the various image, audio, and video generation solutions have now become incredibly good. Even the imitation of people, objects, and structures, etc., works extremely well in many cases now. The use often requires several iterations because control over the results is not fully given. But you can hardly distinguish the AI-generated results from those created traditionally. Also, very well developed are AI-supported software developer tools like Google AI Studio or Lovable. These tools are already fundamentally changing how software solutions are created today and bring enormous possibilities, even for people who actually cannot program.

Specialised solutions in the engineering sector, such as the aforementioned CAD automation solutions, are not yet so advanced. It's not just about text or creative content generation, which AI is already great at. So, for specialised tasks with precise requirements that go beyond text, there is definitely a need for development.

neue verpackung: Perhaps the most important question companies should ask themselves about AI: where do we start? What do you recommend to mechanical engineers as the first step towards AI-supported engineering?

Eckertz: The first step, in my view, should always be a concrete potential analysis within the company. It's not very useful to start with a particular AI technology without knowing exactly where the biggest challenges lie today. Companies should therefore first look at their current situation: where do high efforts, media disruptions or long lead times occur today? Where is a lot of knowledge manually searched for, coordinated or repeatedly rebuilt? Based on this, specific use cases can be identified where AI can actually deliver added value.

It is important to realistically weigh up the benefits and efforts. In addition to the expected added value, topics such as data availability, data quality, and the technical implementation effort also play a central role. This results in a clear prioritisation of which use cases should be tackled first and which may follow later.

neue verpackung: What prerequisites - technological, organisational, or cultural - must companies create to successfully use AI in engineering?

Eckertz: One of the most important prerequisites, in my view, is openness. Companies must be willing to try out new technologies and give them a real chance, even if there are initial acceptance difficulties or not everything works perfectly right away. AI cannot be evaluated purely theoretically; it must be tested in the concrete work context.

At the same time, clear framework conditions are needed. It must be defined which AI solutions may be used, which data can be used, and where limits lie, depending on providers, security requirements, and regulatory guidelines. This clarity creates trust and facilitates everyday use. However, the decisive factor remains the data basis. Without well-structured, available, and maintained data, AI cannot deliver added value. Building and continuously maintaining this data foundation is not a one-time task but must be permanently anchored in engineering. This affects processes, responsibilities, and ultimately also the culture in the company.

And fundamentally, it is of course important to involve employees early and position AI as support, not as a replacement. Where expertise and AI sensibly interact, the greatest benefit ultimately arises.

“AI will change everything”

At the Packaging Machinery Conference 2025, Ralf Schubert demonstrated using his company as an example, how artificial intelligence can be specifically used in packaging machine construction - for example, in robotics, motion planning, and document search. A practical insight into the possibilities of a technology that is changing the world at an unprecedented speed.

neue verpackung: If we dare to take a famous look into the future: What will a typical working day of a designer or development engineer look like in five years?

Eckertz: The working day of an engineer will be more dialogue-oriented in five years, that is, in dialogue with AI. It will be typical for AI to prepare several solution variants, designs, or scenarios based on existing data. The actual engineering work then consists of evaluating these suggestions, classifying them, and making informed decisions. Humans will then work together with AI. At the same time, collaboration across different phases will become closer: requirements, design, commissioning, and service will interlock more because information is more consistently available. AI helps to make connections visible and to show the impacts of decisions at an early stage.

Overall, the working day will be less characterised by routine because it will be better supported. AI will become a constant sparring partner in everyday engineering, but responsibility, technical understanding, and the final decision will remain with humans.

neue verpackung: What impact will this development have on the training of future generations of mechanical engineers?

Eckertz: I believe that the understanding of training in engineering will change significantly. The classic linear model, where knowledge is imparted and applied for as long as possible, fits less and less with a world where technologies and tools are evolving very quickly. In the future, it will be more about imparting fundamental skills: technical system understanding, analytical thinking, problem-solving skills, and the ability to quickly familiarise oneself with new topics. Specific expertise in individual areas remains important but loses its half-life more quickly. Continuous learning will then become much more important.

Engineers must learn to continuously update their knowledge and to integrate new tools, especially AI, into their work in a targeted way. Dealing with AI means, above all, critically questioning results, examining assumptions, and making informed decisions.

Overall, the focus of education is shifting away from pure knowledge transfer towards the development of skills. It is precisely this ability to navigate confidently in a dynamic technological environment that will be crucial in the future.

AI at the Packaging Machinery Conference

Our interview partner Dr. Daniel Eckertz is also a speaker at the Packaging Machinery Conference 2026, which takes place on 16 and 17 June in Nuremberg. Here, he delves deeper into the question of how AI is changing engineering in mechanical engineering.

You can find the full programme and many more details about the event at https://www.packaging-machinery-conference.de/.