Building smarter credit risk and lending solutions
Carrington Labs builds explainable credit risk models trained on your own data, products, and observed outcomes, not a generic score built for someone else's book.
We call this custom Models-as-a-Service: credit risk models you can use across underwriting, offer setting, servicing, and portfolio monitoring, without replacing your existing systems or handing over decision control.
Every model is built around your product, your customers, and your observed outcomes, then continuously refined as your book evolves.
Built around your book
Every model is trained on your product, your customers, and your observed outcomes. Not a population that looks like yours. Yours.
You keep control, always
We provide the model and the explainable outputs. Your team keeps control of policy, cutoffs, and every lending decision.
Continuously refined, never static
A model we build doesn't stop working the day it ships. We keep tuning it against your live performance, not a one-off project that goes stale.
WHY WE STARTED
Our origins
Traditional credit scores were built for a market that no longer exists. Developed in 1989, they weren't designed for today's lending products or borrower profiles.
The data available to assess credit risk has changed too, yet most scores still treat every lender's book the same way.
Carrington Labs was founded on a simple conviction: if every lender's book is different, their credit risk model should be too. We don't hand over a model and walk away. We build it around your data, then keep refining it as your book evolves - the difference between a model and a one-off consulting project.
WHY CARRINGTON LABs
Leading the way in modern lending
At Carrington Labs, we build custom credit risk models around each lender's own book, not a generic score built for someone else's. Our models are built and refined by our team of in-house data scientists and engineers, and every model keeps improving against your live performance long after it ships.
The way we use data and explainable AI to build our credit risk scoring and lending solutions that have been validated is what sets us apart.
Proven with real lenders
Kiva, a US nonprofit lender, saw Early Arrears fall 20% year-on-year in Q1 2025 and 58% year-on-year in Q2 2025 after adopting Carrington Labs.
Live on Snowflake Marketplace
Cashflow Score is available today as a native app on the Snowflake Marketplace for Australian lenders, with more regions to follow.
Built to get sharper, not stale
Every model is monitored and refined against live performance, so it keeps improving with your book instead of sitting still after go-live.
How we work with lenders
Carrington Labs is a credit risk analytics and modelling partner, not a decisioning platform. Here's exactly where the lines sit.
Does this replace our decision engine?
No. Carrington Labs provides model outputs and explainable drivers of credit risk. Your decision engine or policy framework continues to execute every decision.
Who keeps control of policy and cutoffs?
You do. Approvals, declines, referrals, and customer treatment stay with your team, always.
What is Model-as-a-Service?
A credit risk model trained on your own data and outcomes, delivered as a service into your existing workflow. Not a generic score, and not a new platform to run.
who we work with
Collaborating with leaders in lending
We’re proud to work with lenders, fintechs and technology partners who share our commitment to smarter, data-driven credit risk.
FAQs
What does Carrington Labs actually do?
Carrington Labs builds custom, explainable credit risk models for lenders, banks, credit unions, and fintechs. Each model is trained on the lender's own data and observed outcomes, and delivered into their existing systems via API or batch.
Is Carrington Labs a lending platform or decisioning engine?
No. Carrington Labs is a model and credit risk analytics layer. It provides model outputs and explainable drivers; the lender's own decision engine or policy framework continues to execute every decision.
Who owns Carrington Labs?
Carrington Labs is part of Beforepay Group (ASX: B4P), an ASX-listed company that also operates Beforepay, a consumer lending business in Australia.
What is Model-as-a-Service?
It means Carrington Labs builds and maintains a credit risk model trained on your own data, delivered as a service into your existing workflow, rather than selling a generic score or a new platform to run.
Does Carrington Labs require transaction data?
No. Models can be built on bureau, application, or internal performance data, transaction data, or any suitable combination. Transaction data is helpful but not required.
How is Carrington Labs different from a credit bureau or a standardized score?
A bureau score is built for broad use across many lenders. A Carrington Labs model is trained specifically on your product, your customers, and your observed outcomes, so it can separate risk more precisely for your own book.
.png)
.png){kind=logo}
