We wanted to build MES-light: transparency in operation without the complexity of a full MES. Quickly we saw the hardest part was not the dashboard — it was connecting machine signals and OPC UA nodes from different manufacturers. Every machine speaks its own language. Every manufacturer delivers different structures. Integration and maintenance effort grew before the product itself created value.
Instead of solving that with more configuration, we tested how connectivity could be simplified. Less manual mapping, less “model everything first, then see if it works”. The focus shifted: not another surface, but a smarter way to work with raw signals.
That led to an algorithm that could initially do one thing: assign and analyze nodes — recognize patterns in incoming signals without someone pre-defining every variable. Not a finished product yet, but a first breakthrough: the machine was not only connected, it became readable.
That step showed us the potential of AI in industry — not as a generic cloud model, but as a tool that draws structure from real operating signals. We kept researching: what happens when the engine does not only map, but understands what is changing in operation?
Today that has grown into a cognitive engine: In our internal validation it can classify signals without prior mapping, structure them into domains, and make them explainable for people — read-only, with no intervention in the control system. Copass is the product that emerged. elvnode is the company building it.
For us this is not a marketing legend, but an honest path: from a shop-floor connectivity problem to a system that makes machines understandable. That is where we work today — seeking test customers and manufacturers for the first field test with Copass.