Compare software against one real workflow, its inputs and its acceptance evidence before broadening the scope.
Start with a deliverable
Choose one task the team can finish and review: a document prepared from an agreed vehicle record, a regulatory change assessment, or a production list with VIN states. List the users, source systems and expected output. A broad feature list cannot answer whether your own data will work.
Separate the responsibilities
The manufacturer supplies and owns the technical data. The software team maps the fields and explains the checks. The manufacturer’s designated reviewer decides whether the output meets the agreed task. Where an authority, signature provider or receiving system is involved, record that dependency and its separate acceptance criteria.
Build the trial around exceptions
Use synthetic or authorised redacted data. Include a missing mandatory field, an inconsistent value and a corrected record. Check how issues are displayed, who can resolve them and whether the corrected output can be traced to its source. Include export and language needs, as well as user access.
Agree the decision before the demo
Write down what passes, what stops the pilot and who signs off. Compare the reviewed output with the agreed source record. Record unresolved questions and the next trial scope. An attractive demo or a correct file format does not by itself demonstrate authority acceptance.
Illustrative scenario · synthetic data
| Pilot item | Review evidence |
|---|---|
| Input | Synthetic vehicle record A; one missing field |
| Trial | Check → correction → regenerated output |
| Decision | Named reviewer compares output with source; open issues retained |