How should a buyer check grounding and disclosure risk in a generated sample?

**Is the data real?**

No. MirrorFirm's commercial output is generated and synthetic, informed by direct observation of how businesses operate. It is not a copy of real business records.

**Does that remove the need to check?**

No. Generation does not by itself establish that the output is free of resemblance to real people or organisations, and it does not amount to a privacy or regulatory approval. Treat disclosure as a risk to inspect.

**What can a buyer do?**

- Search names, identifiers, addresses and phone formats in the sample against your own reference lists. - Read free-text fields such as handover notes and case descriptions, where specific detail tends to slip in. - Ask how entities and policies were produced and what constraints applied. - Test for repetition, where the same unusual detail recurs across records. - Have your legal or privacy team review the sample under your own standards.

**Is realism a different question?**

Yes. Grounding asks whether the workflow behaviour is plausible. Disclosure asks whether the material resembles something it should not. Review both, separately.

**What should be agreed in a pilot?**

The checks you will run, the sample size you will review and what result would lead you to change the scope.