IT/OT Manufacturing Summit Recap Europe

At this year’s summit, discussions with European manufacturing managers showed a clear shift: the focus is moving away from experimental technology toward solving concrete production problems.

Below are the four challenges we discussed at the event, along with their practical impact on your daily operations.

1. Simple Software Architectures That Keep Production Moving

When connecting ERP, MES, SCADA, and cloud platforms, discussions quickly move from IT systems to functional responsibilities.

The Main Point: The ultimate goal of any system architecture is simple: “Products out of the door matter.”

The Practical Value: Software systems must support physical production, not create administrative delays. Setting clear boundaries between systems prevents IT issues from stopping the assembly line.

World Café on scalable IT/OT architecture in manufacturing. — World Café results on IT/OT architecture, system boundaries and responsibilities across ERP, MES and SCADA.

2. Solving Production Problems Before Implementing AI

Implementing artificial intelligence in a factory fails if the underlying infrastructure and data are not ready, or if the staff lacks the necessary skills.

The Main Point: Introducing AI is primarily an organizational challenge, not a mathematical one. The most reliable approach is to identify a specific shopfloor bottleneck, prove the solution works on a small scale, and only then expand. Don’t start projects you can't scale.

The Practical Value: Avoiding expensive IT failures requires starting with the data and machinery already on your shop floor. Make sure you have the right data for your actual business need, not look for magical improvement ideas in existing data sets.

World Café on scaling Industrial AI in manufacturing operations. — World Café results on Industrial AI, data quality, infrastructure and scalable AI applications in manufacturing operations.

3. Agreeing on Data Meanings Before Connecting Systems

Connecting different software programs is useless if departments and machines interpret the same terms differently.

The Main Point: Data integration fails without clear, shared definitions. Implementation must follow a strict order:

Operational Definition → Data Structure → Software Systems

The Practical Value: Standardizing basic terms—such as what counts as a "cycle," "downtime," or a "good part"—prevents errors from spreading. This ensures that when systems connect, they share reliable information rather than spreading errors faster.

4. Setting Guardrails for Autonomous AI Systems

As AI systems take a more active role in production, the challenge shifts from generating insights to deciding what AI is allowed to do autonomously.

The Main Point: Autonomous AI needs clear guardrails. Manufacturers must define which decisions AI can make independently, when human approval is required, and which safety, quality, and process limits must never be exceeded.

The Practical Value: Clear decision boundaries allow manufacturers to increase autonomy without losing control. AI can react to deviations and optimize processes within predefined limits, while critical or uncertain situations are escalated to human experts. This creates a path toward autonomous production that is controlled, traceable, and safe.

World Café on Data Governance and Digital Twins in manufacturing. — World Café results on data governance, data quality, trust and secure Digital Twins across manufacturing environments.

Aligning Software with Shop Floor Reality

For STIWA Shopfloor Software, these discussions reinforce a key principle: digitalization, AI, and optimization deliver value only when grounded in production reality.

This means building on reliable, well-structured shop floor data, addressing clearly defined production challenges, and designing software around real manufacturing processes and constraints.

Technology creates potential. Connecting it to shop floor reality turns that potential into measurable production value.

Practical Resources for Your Shop Floor

Building a reliable data foundation relies on proven methods rather than theories. Explore these resources to see how these principles are applied in actual manufacturing environments: