How do I check grounding and disclosure risk in generated data?

**Q: Is generated data automatically safe to use?**

No. Generation reduces reliance on real records, but it does not guarantee anonymity or compliance. Inspect it as you would any dataset that enters your environment, and involve your privacy and legal teams.

**Q: How do I check grounding?**

- Ask how the process patterns, such as approval routes, exception types and handover points, were informed by observation of how businesses operate. - Have someone who has run this workflow read a sample of cases and mark which feel wrong. - Compare policy rules in the data with the rules you actually run. Differences are fine if they are stated and intentional.

**Q: How do I check disclosure risk?**

- Search samples for identifiers, names or free text that look like real people or organisations. - Look for rare field combinations that could single out one real case. - Ask what controls applied during generation and what review happened before release. - Ask what you may do if you find something concerning, and who to tell.

**Q: Does Aster Financial help?**

Aster Financial is MirrorFirm's sample organisation, so it can serve as a reference for examples. It is not evidence about your own business.

Record what you checked and the result. That record is useful for your own governance, whatever the outcome.