Condition monitoring of machines is a building block for efficient value chains. The IGF project "AgAVE" has developed methods for setting up assistance systems for process analysis along production chains in Industry 4.0 environments. It uses process-related ""local and a global assistance system(s)"". The local assistants are associated with individual machine modules. They learn and analyse the production processes using Artificial Intelligence (AI) methods (neural networks etc.). The results from the local assistants are evaluated by the ""global assistance system"". It has a central overview of the entire production chain. In the event of an error, it thus provides the operator with information about the error and its possible cause. In some cases it might be that the cause of the error is not based on the machine module that reports it but can be located in a previous production step. This behaviour of diagnostic assistants helps to prevent long production downtimes due to cause analyses. For the communication along the analysis chain, the Industry 4.0 compliant OPC UA protocol is used. Each assistance system carries an asset administration shell to share information along the network. The challenges of setting up and integrating diagnostic assistants are demonstrated using data from machine modules of a production line for packaging material. The structure of the asset administration shell and its means of communicating via OPC UA will also be explained.