Does Carrington Labs make the credit decision?
No. Carrington Labs supplies model outputs; the lender retains policy and final decision control.
What is a custom credit risk model?
A model trained and calibrated on your own portfolio, products, and performance data to estimate probability of default and rank risk — not a generic market score.
What data can the model use?
Any combination, or any single source: transaction data, bureau data, application data, or your internal performance data. It doesn't require transaction data — but turning transaction data into predictive signal is one of our particular strengths.
Does the model work with customer-consented transaction data?
Yes. It works with customer-permissioned (open banking) transaction data, used with consent and delivered as explainable, decision-ready signals.
Is this cash flow underwriting?
Yes — and it's one of the strongest expressions of it. Where a lender has transaction data, we turn it into advanced behavioral signals that predict credit risk for their specific product and portfolio, rather than a generic cash-flow rule set. And if a lender doesn't have transaction data, that's fine — the model can be built on bureau, application, or internal performance data instead. Either way it's calibrated to your own book.
What makes a Carrington Labs custom model different?
The edge is in the modeling, not just the data: advanced feature generation and behavioral features engineered from raw data, calibrated to your portfolio and products, with explainable drivers built in.
How is it different from a generic bureau score?
A bureau score reflects broad-market patterns; a custom model is built around your borrower mix, risk appetite, and product economics, so it separates risk more precisely where your decisions actually happen.
Can we validate it on our own portfolio before going live?
Yes. It's tested against your historical outcomes so you can see performance and impact before deployment.
Does it replace our decisioning platform?
No. It feeds an explainable risk signal into your existing rules, scorecards, and decision engine.
How is the Credit Risk Model different from Cashflow Score?
Cashflow Score is a fast, pre-tuned score built on transaction data — a ready-made signal you can plug in quickly. The Credit Risk Model goes further: it's personalized to your portfolio, products, and objectives, and can be built on transaction data, bureau, application, and internal performance data, or any mix. It's the stronger option when you want a model tuned to your own book.
What if we don't have transaction data?
That's fine. The Credit Risk Model can be built entirely on bureau, application, and internal performance data. Transaction data adds a powerful behavioral signal when it's available, but it isn't required.
.svg)
