Is mismatch resolution just data cleaning with extra steps?

**Is this just data cleaning?** No. Data cleaning normalises a dataset in bulk, usually by rule. Mismatch work is case-based: two live records disagree, and a person must find out why, decide which one is wrong or whether both are right, and change business state through an approved route.

**What makes it hard for a model?** The evidence is spread across systems, partial, and sometimes misleading. A model has to decide what to look at next, tell a timing difference from an error, and recognise when a fix is outside its authority.

**Is the answer always a correction?** No. Outcomes include correcting a record, waiting for a pending item to clear, raising a request with another team, writing off a small variance under policy, or taking no action. A task pack can mark each of these as the expected outcome for different cases.

**Why not rely on rule-based matching tools?** Matching engines can clear straightforward items. The residual items they could not match are where judgement sits, and that residue is what a dataset on this workflow targets.

**Is the data real?** No. MirrorFirm's output is generated. A buyer should still inspect how closely cases follow plausible operating practice and whether any generated detail looks too close to a specific organisation.