Valuation & underwriting copilot estimate

Context

An estimate for a property valuation and underwriting copilot. The trap is the scope change: 'add a new market' sounds trivial but re-triggers comps/market data-readiness and eval — the same phase that dominated the original estimate.

The decision

Make the change-order math visible: a new market isn't a config toggle, it re-runs data-readiness and eval, so estimate the change explicitly and let the margin math show why absorbing it silently is the wrong call.

What most miss

Stakeholders read 'new market' as small because the model is unchanged. The cost is data — comps, rent-rolls, local rules — so the change re-triggers the most expensive phase, not the cheapest.

Stakes

Absorb 'just one more market' silently and margin bleeds out one data-readiness re-run at a time.