How should a lab scope and license generated mismatch material?

Scoping starts with the workflow boundary, not the volume. Decide which systems disagree, which roles act, which causes matter, and whether you need records, tasks, trajectories or an interactive environment. A narrow, well-specified slice, such as one domain and one pair of systems, is easier to inspect than a broad one.

Several limits apply to everything MirrorFirm provides. The material is generated and synthetic. It is informed by how businesses operate, but it is not a copy of any organisation's records, and MirrorFirm does not supply real business records. No release guarantees anonymity, privacy compliance or regulatory approval, so a buyer's own legal and risk review still applies, including a look at grounding and disclosure risk in the sample.

Licence terms are agreed per engagement. Points to raise include whether use is for training, evaluation or both, whether redistribution is permitted, and whether evaluation sets must be kept out of training. This entry does not state terms, prices or timelines; those are settled in discussion with MirrorFirm.

A useful first step is a small sample or pilot scoped around your own acceptance checks: the root cause taxonomy, the policies, the format and the grader. Review it, then decide whether to widen scope.