A new product has no defaults of its own.
The outcome gap: what a lender decides on while a new product waits for its first matured loans.
The gap
A fitted scorecard needs matured loans, and a new product, market or segment has none yet. Until a vintage matures, decisions run on rules borrowed from another book, and the cost of a wrong rule shows only when the defaults arrive.
What fills the months
Most lenders bridge them with policy rules and judgement taken from a neighbouring product. There is nothing else to fit on, so this is not a mistake. But borrowed rules carry the other book's assumptions — its applicants, its channel, its prices — and nobody can say in advance how far those assumptions travel.
A scorecard before any repayment
On launch day, Day-Zero turns the product's data dictionary and feature statistics into a closed-form scorecard: points, grades and reason codes, with a written reason for every weight. No outcome labels go in, and the validator can read the card before the product goes live.
Every matured loan, learned in context
As loans mature, each one becomes part of zypl.Foundation's context. The model scores beside the current process, with no training run for each product, and the book it learns from is the lender's own.
Service moves on your sign-off
The model takes over when the lender's own held-out outcomes show it ranks better and the risk owner signs. Until then it scores in shadow, and if the evidence turns, service moves back.
Where to begin
Send the product's data dictionary and feature statistics. A Day-Zero Readout comes back within 48 hours with the top drivers, the expected ranking band and the data-readiness gaps. The scorecard and the method stay closed.