From Automation to Artificial Intelligence

Rösberg Engineering is a family business in its third generation – and one that does not rest on tradition. With around 200 employees, the Karlsruhe-based company has been providing automation solutions for the process industry – including chemicals, pharmaceuticals, oil, and gas – for over 60 years.
However, while many mid-sized companies are still contemplating how to integrate AI, Rösberg has already begun to forge its own paths.

Philipp Parker, AI Lead and Innovation Driver at Rösberg, describes his role as "an interface between innovation, AI, and software development." In the so-called R-Lab, a protected innovation space, employees can contribute, experiment, and further develop new ideas. "From the very beginning, we wanted innovation to come not just from the top but from the middle of the organization," Parker says.

From the ChatGPT Moment to a Company-wide AI Strategy

As in many companies, Rösberg's journey also began with ChatGPT. "Before 2022, AI was not an operational topic for us," Parker admits openly. But when OpenAI released ChatGPT in November 2022, it was the trigger: "Within a few months, we had our first AI project in the R-Lab."

Back then, Parker recalls, it was the "Wild West." Colleagues were testing ChatGPT privately, writing poems, experimenting – and raising the first questions about data protection and corporate use.
"We wanted to harness that energy but also alleviate the uncertainty," says Parker.

The result: the “Rösberg Assistant” – an internal chatbot based on modern language models, integrated into Microsoft Teams, managed by IT, and with a clear governance structure. "You only gained access after completing a training and usage agreement – that was important to us."

Thus, AI did not become a chaotic topic but rather a structured learning journey for the workforce.

From Experiments to Tangible Benefits

After a year, Rösberg took stock: In an internal survey of nearly 100 users, there was an average time savings of 54 minutes per week due to AI support – with a range from zero to over four hours.
"That sounds trivial, but over a year and 100 employees, it's a massive productivity increase," Parker explains.

A concrete example: In the company’s own software Liveforms, engineers can create inspection sheets and checklists. This used to happen manually via a drag-and-drop editor. Today, an AI assistant automatically generates the forms based on a brief description – saving over 30% of the time in the creation process.

"The real added value comes when AI is targeted and meaningful – not as an end in itself, but as a tool that truly facilitates work," Parker states.

Innovation Requires Safe Spaces – and Management Support

What sets Rösberg apart from many others is that the company has created a structured culture of innovation.
Every employee can submit ideas – simply via a Teams form or email. The R-Lab team evaluates, supports, and accompanies the implementation.

But Parker emphasizes: "This only works if everyone pulls together – innovation team, IT, and management. If only one department marches, it leads to tensions."
His motto is therefore: "Don’t prohibit, but support." AI must be allowed to be tested within a protected framework.

Small Steps, Big Impact

Rösberg follows a pragmatic approach: "Start small – but start right."
Instead of immediately opting for large, complex AI systems, the team deliberately started with small use cases to build acceptance and trust.

"Prototypes are quickly built – but implementing something that is clean, secure, and scalable is the true challenge," Parker explains.
That's why the company is currently focusing on building its own AI infrastructure that enables the long-term integration of models directly into its own software – "behind buttons and functions, where the user works."

Between Cloud and Edge – the Future is Hybrid

Technologically, Rösberg is open – but not naive. "Cloud is certainly powerful when it comes to scaling. But sovereignty and data ownership are equally important," says Parker.
In his view, smaller, local models will play a bigger role in the future: "You don't need a huge model that can write poems if all you want is to process device information."

He himself experiments privately with local models on a Mac Mini. "Technology is becoming more efficient, cheaper, and more accessible. I believe we will see many specialized small models in the future."

Outlook: The Future of Engineering

When Parker looks ten years into the future, he sees engineers working with AR glasses and voice co-pilots:
"The systems will combine context from sensors, ERP, and software, display information, and provide actionable recommendations. Much will no longer be searched for or manually created, but rather validated."

The vision: A voice-controlled, visual assistant that bridges the gap between humans, machines, and data in an industrial environment – intelligent, secure, and hands-on.

Conclusion: Courage, Structure, and European Strength

For Parker, it is clear: AI is a key technology, but not an end in itself.
"If we combine our technical expertise, precision, and quality with AI, it can create a real competitive advantage for Europe."

What is needed for that?
"More optimism, courage, and collaboration. We have outstanding research, strong companies, and smart minds in Europe. If we can bring that together, then we are not out of the AI race by a long shot."