use cases

Improve decisions for borderline and referred applicants

Add stronger credit risk evidence where your existing scorecard or policy has the least certainty.
Talk to us
Explore the Credit Risk Model

The problem

The clearest approvals and declines are rarely the hardest part of underwriting. The difficult decisions sit near the policy boundary: applications referred for manual review, applicants declined by a narrow margin, thin-file customers, and borrowers whose current financial position is not fully represented in their traditional credit data. Broad score bands can place materially different borrowers in the same risk segment, and manual reviewers then rely on incomplete information or inconsistent overrides. The objective is not to loosen credit policy; it is to improve risk separation where the existing process is least decisive.

How Carrington Labs can help

Improve risk separation near your cutoff
Develop a lender-specific Credit Risk Model that estimates probability of default using the data relevant to your product and portfolio.
Add a current cash flow signal
Where customer-permissioned transaction data is available, add current income, liquidity, obligation, and stability signals to supplement your existing view.
Structure second-look reviews
Apply additional risk evidence to referred or narrowly declined applications, and test before changing policy.
Talk to us

How this fits in your workflow

Magnifying glass on an ID.
Application underwriting
Use the model output alongside bureau data, application information, existing scorecards, and policy rules.
A Computer.
Referral routing and review
Prioritize the applications where more information is most likely to change the decision, with consistent risk estimates and drivers.
A Person giving a pitch infront of a graph.
Second-look underwriting
Reassess a defined population of applicants who narrowly missed the original approval criteria.

Who is this for

Carrington Labs supports lenders aiming to responsibly expand access.
  • Digital lenders
  • Issuers
  • Fintechs
  • Financial Institutions
Talk to us

What Carrington Labs provides

Inputs it can use
Application, bureau, and internal performance data, plus customer-permissioned transaction data where available. Transaction data is not required.
Outputs it returns
Probability-of-default estimates, risk scores or segments, borrower-level drivers, and referral or second-look strategy simulations, via API or batch.
You keep control
Carrington Labs provides analytics and model outputs. The lender retains control over eligibility, policy, approvals, declines, referrals, and customer treatment.

FAQs

What is borderline applicant underwriting?
Borderline applicant underwriting focuses on applications near an approval, decline, or referral threshold where the lender's existing data or score does not provide sufficient separation to make a confident decision.
What is second-look underwriting?
Second-look underwriting reassesses a defined group of declined or referred applications using additional risk evidence. It is not a blanket override of the lender's original policy.
Does Carrington Labs replace our current scorecard?
No. Carrington Labs can add a lender-specific model or supplementary cash flow signal alongside your existing scorecard, bureau data, and policy framework.
Is this only for thin-file borrowers?
No. It can also help distinguish applicants with established credit files who have similar traditional scores but different current financial behavior or product-specific risk.
Can we test the model before changing policy?
Yes. The model can be evaluated against historical portfolio outcomes, subject to data quality and sample size. Lenders can then simulate how alternative cutoffs or referral strategies might have performed.
Does Carrington Labs make the approval decision?
No. Carrington Labs supplies risk analytics and model outputs. The lender retains control over policy and final decisions.

Related resources

Credit Risk Models and Analytics
Guide & Explainer

5 Signs Your Lending Team Has Outgrown Its Current Credit Model

Five early signs a lender's credit risk model no longer fits their growing loan book, and what to do before it becomes a bigger problem.

Read Article
Custom score models improve risk separation for near-prime lenders
Credit Risk Models and Analytics

Why Near-Prime Lenders Need Custom Score Models for Better Risk Separation

Near-prime borrowers vary widely within one score band. See how custom, lender-specific score models turn existing bureau, application, or transaction data into better risk separation.

Read Article
Lender-specific credit risk modelling services for lenders
Credit Risk Models and Analytics
Guide & Explainer

Credit Risk Modelling Services: Why Lender-Specific Models Matter

Credit risk modelling services help lenders build lender-specific models that improve approval quality, pricing precision, margin, and portfolio performance.

Read Article
Credit Risk Models and Analytics
Guide & Explainer

What Is a Custom Credit Risk Model as a Service?

A custom credit risk model as a service explained: what it is, how it differs from generic scores, and what to look for in a provider.

Read Article

Explore solutions

Cashflow Score
Credit Risk Model
Credit Offer Engine
Financial Health Summary
Cashflow Servicing

See how Carrington Labs

fits into your lending process.

REQUEST A DEMO

Solutions

Cashflow scorecredit risk modelCashflow ServicingCredit offer ENGINEfinancial health summary

Resources

AboutCareersBlogWhite PapersmarketplaceCOMPAREprivacy policy

Contact

book a callLinkedin
Carrington Labs is a registered business of Beforepay Group Limited, Suite 9.01, 77 Castlereagh Street, Sydney NSW 2000. Copyright © 2026 Beforepay Group Ltd, ABN: 63 633 925 505. Terms and conditions apply. New customers are subject to eligibility criteria.

This service is not currently offered in  China, Russia or Taiwan. If you are interested in our solution, please contact us.

Beforepay Group Limited and its Beforepay goods and services are in no way sponsored, approved or affiliated with Quadpay Inc., Openpay Pty Ltd, Afterpay Pty Ltd, ZipMoney Payments Pty Ltd and their associated entities/affiliated companies.