A production model is a model actively deployed within an institution's live underwriting pipeline, generating outcome-probability or duration predictions that real capital decisions are actually priced against, as distinct from an experimental model still in development, a retired model no longer in active use, or a reference artifact that supports the modeling operation without generating predictions itself. Only production models should be counted toward a claim about an institution's active predictive capability, since a broader registry count that includes experimental and retired entries alongside currently deployed models describes the registry's total history rather than the institution's current, working fleet. A production model should also carry its own calibration record, versioned and re-validated on a defined cadence, because a model's production status is not permanent: a production model that fails a scheduled re-validation should be restricted or retired from production use, moving it out of the production category until it either passes retraining or is formally retired, a movement that should itself be recorded as a dated event in the registry rather than handled quietly off the books, so a later reviewer can reconstruct exactly when the model stopped being relied on for real decisions and why that determination was made.
Working through a diligence process?
Institutional partners evaluating a position against this platform's outcome and duration models are welcome to reach out directly.
