ZEISS Microscopy GmbH develops premium-segment microscopes and, together with doubleSlash Net-Business GmbH, transitioned its field service to a predictable and data-driven approach. The objective was to establish a reliable foundation for condition analyses and to support technicians during on-site operations. This is based on uniform device connectivity, standardized data, and a robust historical record.
The challenge: digitally consolidating operational data and service knowledge as a basis for predictive maintenance
ZEISS Microscopy develops highly complex microscope systems in the six-figure price range that are deployed worldwide. These systems generate large amounts of operating and condition data—but in different formats, depending on the device type. A uniform database that provides data securely and in a standardized format was therefore necessary for the further development of the service offering.
In addition: the devices operate extremely reliably. Failures occur rarely, meaning that initially only a small number of events were available as training data for AI models. For the target vision of “predictive maintenance,” a valid history first had to be built up and the selection of truly relevant parameters clarified—so that subsequent analyses would be targeted and reliable.
Furthermore, the various device types had to be connected using a shared architecture and a uniform security concept. Only this harmonization enables standardized classification of conditions as well as subsequent scaling to additional product lines.
The solution: uniform connectivity, a clean data foundation, and AI-supported service assistance
At the start of the project in 2018, the target visions and use cases were refined together with doubleSlash. Moderated workshops resulted in a clear technical architecture that takes into account both device connectivity and the subsequent expansion of data-driven services.
Implementation followed the holistic approach model “Connect – Make Smart – Monetize.” For ZEISS, the focus was on establishing a robust connectivity and data foundation. Security requirements and scalability were considered from the outset, ensuring that the architecture remains extensible in the long term.
After just one to two months, the first microscope device type was connected and made available via a central platform. Over time, additional device types were to be connected and released step by step. Because the devices delivered heterogeneous data, formats were harmonized and contents standardized. Since then, the data has been transmitted securely to a central platform, where it forms a reliable historical record as the basis for analysis, visualization, and service operations.
Central device and update management ensures that software versions remain consistent and that new functions can be rolled out in a controlled manner—an important aspect for international service organizations.
On this basis, the project team developed models for condition analysis and initial predictions. To that end, consistent datasets collected over longer periods were built up, enabling deviations to be reliably detected and interpreted in a technically sound manner.
In addition, a generative service assistant was introduced. It leverages documented experiential knowledge from maintenance sessions and manuals and provides it in the form of a chatbot. This gives technicians step-by-step procedures, tool and material lists, and guidance on critical points—allowing them to act faster and more safely on site, even if they lack deeper experience.
In parallel, service processes in international operations were professionalized. Central provision of knowledge and data creates globally comparable workflows—regardless of the individual experience level of specific technicians. This enables service to be scaled step by step and new product lines to be connected efficiently.
The result: more predictable service, faster diagnosis, and predictive maintenance
With the connectivity established and a clean data foundation in place, service operations at ZEISS Microscopy can act far more predictably. Service teams can now identify failure risks early, proactively inform customers, and bundle maintenance appointments in a targeted manner. Fault diagnosis is simplified because information and recommended actions are available centrally. The generative assistant provides step-by-step procedures, supplemented with tool and material lists, highlights critical points, and thus supports less experienced technicians in the field as well.
Based on the standardized historical record, robust models for condition analyses and predictions were developed. Predictive maintenance was implemented step by step after the necessary data history had been built up. Service operations work more efficiently and can respond faster—and ZEISS strengthens customer loyalty through proactive, data-driven support in the field.