Artificial intelligence is no longer a future topic in the German mid-sized companies. Many companies are testing initial use cases, establishing internal structures, and looking for ways that AI can create real value in processes, products, and decisions. This is exactly what Dr. Marius Biederbach and Maximilian Hahnenkamp discuss in their podcast – practical, unvarnished, and with a clear focus on what really matters for companies: impact instead of vision, impact instead of strategy slide.

From Doctoral Researcher to AI Responsible in the Mid-sized Sector

Marius Biederbach originally comes from business administration, switched early to business informatics, and dealt intensively with digital ecosystems in research and his doctoral thesis:
How do technologies like AI change established industries? How do traditional companies respond when digital players – from start-ups to Google – enter their value creation?

His central research finding:
Digital technologies do not necessarily threaten established companies – they can even strengthen their position if they engage in innovation cooperations and actively integrate new technologies.

With this background, Biederbach moved into practical work: as Digital Transformation Manager and Lead AI at the internationally operating Stern-Wywiol Group.

Who is the Stern-Wywiol Group – and what does Biederbach do there?

The Stern-Wywiol Group is a global producer of food and feed ingredients with 13 specialized business units – for example, in the field of enzyme mixtures, which ensure that bread succeeds consistently worldwide.

In his role, Biederbach builds the bridge between IT and business units:

• He is responsible for the AI strategy,

• identifies and prioritizes use cases,

• ensures knowledge transfer,

• and ensures that new technologies are introduced responsibly and scalably.

The team is anchored in IT – consciously. Because according to Biederbach, AI and digital transformation are primarily technical topics, which, however, can only generate impact in close coordination with the specialist departments. This connection is one of the keys to success.

How does the mid-sized sector really stand on the topic of AI?

The impression of many: the mid-sized sector is slow, hesitant, or even backward.
Biederbach disagrees:

“The mid-sized sector is faster, more pragmatic, and implementation-oriented than many believe. The greatest hurdle is not the mindset – but the data.”

Many mid-sized companies have clear responsibilities, short decision-making paths, and high efficiency pressure. This facilitates rapid pilot projects and pragmatic experiments.

At the same time, Biederbach identifies three central challenges:

1. Data quality and availability:
   Without clean, accessible data, every AI project remains piecemeal.

2. Structures and responsibilities:
   Who implements AI – IT, specialist department, or a hybrid team?

3. IT security & data protection:
   Especially in dealing with sensitive production and formulation data.

From Thinking to Doing: Why "AI Impact" is more important than "AI Strategy"

While many companies are developing complex three-year roadmaps, Biederbach opts for the opposite:

“Start, learn, and develop the structures while doing so – not the other way around.”

He warns against waiting too long for the “perfect” model, the “perfect” database, or the “perfect” governance document.
Because: Technology is evolving faster than any strategy can depict.

The Biggest Mistakes When Starting with AI

Biederbach identifies three typical stumbling blocks:

1. Waiting too long – for perfect data, perfect tools, or perfect strategies.

2. Too much governance before the first use cases – structures develop best through experience.

3. Treating AI as an IT topic or a purely specialist topic – the solution lies in the middle: collaboration.

This mix of pragmatism and structure is crucial to achieve real impact.

Concrete AI Applications at Stern-Wywiol

1. Individual Internal Chatbots – Developed by Employees

Employees build small chatbots themselves to support them in processes – for instance, when preparing data for SAP Signavio.
An example of how AI increases productivity in everyday life without immediately triggering a large project.

2. The "Product Finder" – A Scalable Use Case

A particularly impactful use case is the Product Finder:
A central AI-supported tool that consolidates product data, formulations, development statuses, and categorizations.

• Sales can respond to customer inquiries faster,

• Product Management gains clarity about product variants,

• R&D benefits from structured expertise.

One of the greatest effects: Knowledge is preserved, instead of disappearing into heads.

How Does One Measure the ROI of AI?

Time savings are the most important measurable factor for Biederbach.
Saved hours can be translated into full costs and strategic benefits.

But at least equally important is the quality improvement – for example, through the safeguarding of know-how or standardized product decisions.

Change Management: AI Ambassadors and AI Coaches

Technically, AI is rarely a problem today. The challenge lies in understanding:

• AI works probabilistically, not deterministically.

• It delivers probabilities, not fixed system responses.

To build up knowledge across the board, the Stern-Wywiol Group has established a network of AI Ambassadors and AI Coaches.

They are in the specialist areas, not in IT, and serve as local experts, multipliers, and sparring partners for use cases.

How Are Use Cases Created and Scaled?

Initially, many use cases arose spontaneously from the organization. Today, the company follows a structured approach:

1. Assess and prioritize by effort & benefit

2. Analyze together with the specialist departments

3. Proactively address processes and offer optimizations

From reactive idea collection, it shifts to proactive enablement, where AI not only solves problems but also reveals new potentials in the process.

What Will AI Look Like in the Future?

Biederbach looks at new developments for 20 minutes daily and is convinced:

• We are only at the beginning.

• LLMs will become more performant and specialized.

• The greatest lever lies in the combination of systems + data + models.

In the long term, AI will not only improve products but also reshape entire business models.

Advice for All Mid-sized Companies

When asked for his most important tip, Biederbach answers without hesitation:

“Start and look at your own data. That is the beginning of everything.”