Challenge
Vehicle-wide requirement documents define requirements for the entire vehicle architecture - with consequences for many hardware and software components. With every new version, component and function owners as well as suppliers had to check by hand which requirements had changed and whether that affected their own scope. At this document size, that is slow and error-prone.
Solution
We built a web application with a cloud backend that understands specifications instead of just storing them:
- Automatic comparison - every new version is matched against the previous one.
- AI summaries - a language model describes what changed in substance, not just which lines.
- Semantic search across several specifications, even where terms are used differently.
- Annotations - requirements can be tagged, for example when they are legally relevant.
- Import pipeline for new releases so the system stays current without manual work.
Result
- One central tool instead of scattered spreadsheet comparisons.
- Change history with semantic summaries and visual diff - generated automatically.
- Noticeable time savings for everyone who has to assess requirements.
- A clean data foundation for further AI use cases.
* Project: we owned the delivery · ● in operation: we run the solution today
