The path to Production Level 4 is paved with many elements
The vision of Production Level 4 (PL4), the major update of Industrie 4.0, assumes that companies in the future will build customized production lines to enable the profitable manufacture of even small batch sizes. On the foreseeable horizon, central tasks will be performed by so-called production bots (software agents). For example, the bots will search digital platforms to put together the necessary processing steps and select the appropriate means of production (hardware or software). While doing this, they can consider various criteria like energy efficiency, production schedules, and transport routes. “Technically, that is all possible, but how does it ultimately make money?” asked Prof. Martin Ruskowski, Chairman of the Executive Board of SmartFactoryKL (SF-KL). Currently, SF-KL is working with partners to develop AI solutions that will be available to SMEs as a service. “To date, no concrete method exists for a company to offer a service or for another company to use it,” said Ruskowski. Transactions could be processed through an annual subscription or invoiced on a case-by-case basis. But so far, these ideas are only proposals. SmartFactoryKL is determined to fill this void.

 

SF-KL leads a consortium at the SME Digital Center Kaiserslautern (MDZ), which advises SMEs and supports their digitalization projects. “We are experiencing strong demand for artificial intelligence services – for example, for quality control – but when asked about the costs, we find it difficult to answer,” said Jonas Metzger, Head of MDZ. He added, “There is just no universally applicable, practical answer on the market.” Consequently, SmartFactoryKL is working on “federated (collaborative) learning” with member organizations. The complexity of this topic was made clear by the experts in a recent SmartFactoryKL Live stream broadcast. The term describes a machine learning technique that trains an algorithm across multiple decentralized edge devices. Each participating device (server) holds its own local data sets, without exchanging them. In this way, company secrets are protected as the local datasets are not uploaded to one server and the AI gets “smarter” with every use. This is exactly what leads to the question regarding cost structure. The fact that AI is learning something new with every order means that it becomes more and more valuable as a product. This would have to be included in a price calculation. Tatjana Legler, project manager for TU Kaiserslautern, clarified this issue by asking these questions: “How much has the customer contributed to the improvement of the AI? How much more does the next customer benefit from the higher quality AI skills and how can these be captured financially?”

 

“Today, we must not only provide the impulse for technology transfer, we must also push the development of innovative business ideas. That is only logical. We want to initiate the discussion, not implement any one idea,” said Ruskowski. The SmartFactoryKL management board has now agreed that the workshops will be seen as important milestones in the step-by-step process of making the Production Level 4 (PL4) vision a practical reality. “PL4 aims for resilience and sustainability. Many are already excited about that,” said Ruskowski. “Now we want to show how our ideas can also be used to make money. That is what convinces many more decision makers.”