MCP at the Sirris demonstrator lab in Kortrijk

Digital production tangible on the shop floor 

At Sirris' demonstrator lab in Kortrijk, you can discover what a modern, paperless production environment looks like in practice. No theoretical models, but a realistic shop floor where digital technologies come together to make production processes more efficient, transparent and people-oriented.

The lab has been set up as a learning environment. You don't have to redesign your entire production immediately: you can draw inspiration from individual concepts, tools or practices and adapt them to your own context. Sirris accompanies this process with specific expertise and practical advice. 

Operator support as a starting point 

The demonstrator focuses heavily on supporting operators in their daily work, using a best-of-breed approach. This means that specialised software solutions work together rather than having one comprehensive system.

The individual applications are linked via a unified namespace (UNS). This means that every action on the shop floor, from production and quality to energy consumption is recorded, enriched with context (what, where, when, unit) and made available centrally. The result is a single consistent data source that is reusable across the organisation.

Moreover, the demonstrator lab continues to evolve: new technologies and applications are continuously added, so that the lab grows along with developments in industrial digitalisation. 

MCP: a new way to query production data 

An MCP (model context protocol) interface was added to the lab recently.

It’s an open standard that describes how AI models can interact securely and in a controlled manner with external systems and enterprise data. In other words, it enables the targeted deployment of AI in an industrial environment in a way that maintains control and transparency.

Furthermore, MCP allows natural language communication with production data from both time-series databases and relational databases, using a large language model (LLM).

Traditional dashboards remain important for fixed KPIs and periodic monitoring. They offer stability and consistency, while MCP adds something new: speed and flexibility. A production supervisor can now ask questions such as:

• What is the most common error at a particular work station?
• Where are differences in efficiency between operators?
• How much energy does each process step consume? 

Within minutes, insights are generated without any intermediate technical steps and action can be taken. 
 

From data to insight, without barriers 

The great added value of this approach is that data becomes accessible to more people. Those with access to the data can also query it in plain language. This lowers the barriers for analysis and speeds up decision-making.

However, there are certain conditions: 

• The right data must be available
• That data must contain sufficient context
• There must be a secure MCP link to the AI model
• And the results must be presented clearly, for example in a report or overview

In this way, production data is not just collected but also actively used for optimisation, training and day-to-day management.

In the demonstrator lab, MCP acts as the bridge between production data and natural language interaction. This gives AI a concrete, practical role in production and supervision.