Why would a model lab want data about credit escalation specifically?

**Why would a model lab want data about credit escalation specifically?**

Because it concentrates several behaviours that general business text rarely shows: knowing the limits of your own authority, deciding when to refer, writing a handover that someone else can act on, and handling a decision that comes back with conditions.

**Is this only about lending?**

No. The same structure appears when an insurance adjuster's settlement exceeds their limit, when a utilities hardship agent needs approval for a write-off, or when a finance clerk needs sign-off for a credit note. Lending is a clear anchor, and a release can vary the domain.

**What does a model learn that it would not learn from policy documents alone?**

Policies state thresholds. Workflow material shows how those thresholds meet messy cases: a request that sits just under a limit, a borrower with several linked facilities whose combined exposure crosses it, or a referral sent to the wrong level and bounced back.

**Is the data real?**

No. MirrorFirm generates synthetic material. It is informed by direct observation of how businesses operate, but the output is generated, and a buyer should still inspect it for grounding and disclosure risk before relying on it.

**Where should a lab start?**

With a small sample covering one domain, so you can judge whether the authority structure and handovers match what your training or evaluation needs.