Stadler is one of the leading manufacturers of rail vehicles. Headquartered in Bussnang, Switzerland, the company builds trains, trams and locomotives that stay in service for 30 years and more. Every customer order is unique, with its own suppliers, specifications and drawings.
Challenge
A train stays in service for 30 years and more. When a part fails years later, often only a serial or material number is known. In after-sales alone, each customer accounts for more than 100,000 documents and millions of structured data records. The answer to a service request has to be pieced together from several systems and IT generations:
- Look up in several archived Infor systems which article and which supplier are behind it.
- Find the customer order and work out which vehicles and series are affected.
- Search SharePoint for the specification, drawing, bill of materials and supplier documents, across more than 30,000 documents per project.
- Derive the maintenance action from the service manual, supplier instructions and expert knowledge.
That took around half a day per service request and engineer. The knowledge existed, but it was spread across decades, systems and formats, and every search started from zero.
Solution
Together with Stadler and JiVS, Genow built the Service Recall Agent. The service engineer asks a question in natural language. The agent breaks it down, queries the material master and purchasing data in the JiVS archive and the project library in SharePoint in parallel, checks every partial answer and sums up the result with sources.
The front headlight on a FLIRT needs replacing. Which article, which supplier, and what do we need to check?
- JiVS material master Article and supplier found
- JiVS purchasing Price history and lead time
- SharePoint project Data sheet, drawing, bill of materials
- Recommendation What to check and replace, every statement sourced
The foundation is the Genow Context Engine. It learns how Stadler speaks and thinks:
- It recognises what every archive record and every document is about.
- It links article, supplier, project and vehicle into one identity across JiVS and SharePoint.
- It knows abbreviations and part names, in German and English.
- It understands how material, installation location, project and drawing connect.
Every query respects the permissions from Entra ID and the source systems. Every step is logged.
Why Genow?
Earlier feasibility studies hit their limit as soon as structured archive data and unstructured documents had to come together in one answer. Genow connects both: the Context Engine links records from JiVS with drawings and specifications from SharePoint, and every statement points to its source. That makes knowledge from retired systems usable again.
“Agentic AI has developed enormously over the last two years. Today you can achieve strong results even with legacy data. Our advice: start small and add further use cases and data step by step.”
How the setup came together
The Service Recall Agent was set up within just 10 weeks: clarify use cases and data access, set up and connect the agents, then test and refine them together with the service team. In November 2026, Stadler, JiVS and Genow showed the Service Recall Agent live at the SAP NOW AI Tour in Zurich.
Results
With Genow, the service engineer gets what the next action needs in seconds: installation locations and affected series, article, supplier and price history, specification, drawing and bill of materials, plus a clear recommendation on what to check and replace.
- Seconds instead of half a day per service request
- Over 80% faster access to information in day-to-day service
- 100,000+ documents per customer and millions of records, searchable in one place
- 10 weeks setup time
- Every statement sourced, with the original one click away
Archived data from retired systems flows back into daily work instead of sitting unused in the archive.
Outlook
Stadler continues to test the solution and is expanding it step by step. Further agents for the material master, purchasing, and warehouse and logistics are planned on the same archive. For the enterprise rollout, SAP data, unstructured documents from JiVS and further customer projects will be added.
Knowledge built up over decades is a competitive advantage for Stadler. With Genow it becomes usable where it is needed: in the answer to the next service request.