Why It Matters
Traditional credit scoring excludes large segments of the population, limiting economic mobility and leaving lenders on the sidelines of potential business. By harnessing bank transaction data and AI, lenders can make fairer, more accurate decisions, opening credit to gig workers, newcomers, and others while maintaining profitability—a timely evolution as digital finance and alternative data become mainstream.
Key Takeaways
- •Traditional credit scores miss thin‑file borrowers, limiting inclusion.
- •Bank transaction data enables cash‑flow underwriting for accurate risk.
- •AI models predict defaults using behavioral spending patterns.
- •Carrington Labs built models inside lending business for data access.
- •Personalized scores improve approvals while keeping lender profitability high.
Pulse Analysis
The fintech landscape is finally confronting the limits of three‑decade‑old credit scoring. Traditional credit bureaus rely on sparse, historical credit activity that excludes thin‑file or gig‑economy borrowers, leaving a large segment of consumers outside affordable financing. By tapping into real‑time bank transaction data, lenders can construct a richer, behavior‑driven view of cash flow, dramatically improving risk assessment and fostering financial inclusion. This shift to alternative data aligns with growing demand for more equitable credit access and reflects broader industry moves toward data‑centric underwriting.
Carrington Labs exemplifies this evolution with its cash‑flow underwriting platform. Leveraging AI and machine‑learning, the company extracts behavioral signals—such as pre‑emptive spending cuts before cash crunches or rapid fund transfers to meet obligations—from billions of transaction records. These features feed personalized credit scores that predict default probability far more accurately than generic bureau scores. The result is a nuanced risk model that can approve borrowers previously declined, while preserving or even enhancing lender profitability through tighter loss forecasts.
Building the analytics engine inside an existing lending operation gave Carrington Labs unparalleled data volume, rapid feedback loops, and direct alignment with client outcomes. This integration enables swift model iteration, compliance‑ready transparency, and the ability to meet regulator expectations for explainable AI. As fintechs and traditional banks seek scalable, inclusive lending solutions, the combination of transaction‑based underwriting, AI‑driven insights, and personalized scoring is rapidly becoming the new standard for credit decisioning.
Episode Description
Cash flow underwriting, explainable AI, and credit risk analytics are changing how lenders approve borrowers and set loan terms. Tedd Huff, CEO of fintech advisory firm Voalyre and founder of Fintech Confidential, sits down with Jamie Twiss, CEO of Carrington Labs, and Kasey Kaplan, Chief Product and Commercial Officer, to break down how behavioral signals in bank transaction data outperform traditional credit scores.
Over 50 percent of loan applicants cannot produce a reliable credit score, leaving self-employed workers, gig earners, and younger borrowers locked out of the system. Carrington Labs uses billions of lines of transaction data to build personalized, explainable machine learning models per lender, per product, and per customer segment. The conversation covers their "control point" approach to AI, lifecycle underwriting beyond origination, open banking friction in the US, and a five-year outlook on embedded, agent-driven lending.
FIND OUT MORE
1️⃣ Map analytics to every step of your lending funnel to find exactly where borrowers drop off and why.
2️⃣ Buy best-of-breed origination and servicing tools instead of building proprietary underwriting tech in-house.
3️⃣ Start with off-the-shelf models, lend small, collect performance signal, then shift to custom models fast.
4️⃣ Offer higher loan limits to borrowers who sync more accounts through open banking.
5️⃣ Track how borrowers respond to financial scarcity and build those behavioral patterns into your credit criteria.
Guest
Jamie Twiss LinkedIn: https://www.linkedin.com/in/james-twiss/
Kasey Kaplan LinkedIn: https://www.linkedin.com/in/kaseykaplan/
Company
Carrington Labs: https://www.carringtonlabs.com/
Carrington Labs LinkedIn: https://www.linkedin.com/company/carringtonlabs/
Beforepay Group: https://www.beforepaygroup.com
Fintech Confidential
Podcast: https://fintechconfidential.com/listen
Notifications: https://fintechconfidential.com/access
LinkedIn: https://www.linkedin.com/company/fintechconfidential
X: https://x.com/FTconfidential
Instagram: https://www.instagram.com/fintechconfidential
Facebook: https://www.facebook.com/fintechconfidential
About the Guests
Jamie Twiss is CEO of Carrington Labs and Beforepay Group. He began his career at McKinsey & Company, held senior banking roles including Chief Data Officer at a major Australian bank, and now leads the development of explainable AI credit risk models for lenders globally.
Kasey Kaplan is Chief Product and Commercial Officer at Carrington Labs. With over 15 years across payments, program management, and fintech lending, he leads commercial execution across credit risk scoring, cash flow underwriting, and loan limit solutions.
About the Company
Carrington Labs is the AI and enterprise software division of ASX-listed Beforepay Group. The company builds explainable AI credit risk scoring, cash flow underwriting, and loan limit solutions for banks and non-bank lenders worldwide, having powered over 4 million loans through its sister business.
About the Host
Tedd Huff, CEO of fintech advisory firm Voalyre and founder of Fintech Confidential. With 25+ years in fintech and payments, he brings entertaining and informative conversations focused on the people, tech, and companies that change how you pay and get paid.
DD3 Media
Fintech Confidential is a production of DD3 Media, a media creation, management, and production company delivering engaging fintech content globally.
Chapters
00:00 Episode Highlights
01:04 Welcome to Fintech Confidential
01:13 DFNS: Wallets as a Service (sponsor)
02:34 Meet Carrington Labs
04:48 Casey FinTech Origin
06:05 Jamie Credit Risk Path
07:59 Mission Beyond Scores
10:16 Cashflow Underwriting
13:35 Alternative Data Behaviors
17:37 Built Inside Beforepay
21:01 AI Control Points
24:07 Deterministic Vs Inference
28:35 Keeping Bias Out
34:48 Real Client Turnaround
36:44 Funnel Friction Signals
38:25 Optimizing Drop Off
39:21 Sky Flow: Building Fast and Secure (sponsor)
40:21 Product Specific Risk Models
42:32 From Shelf To Custom
43:34 Model Retraining Workflow
47:05 Siloed Versus Consortium
48:59 Cashflow Behavior Insights
50:25 Feature Engineering Matters
51:41 Macro Shocks In Data
56:07 Lifecycle Servicing Signals
57:36 Limit Management Uplift
58:55 Open Banking Pushback
01:03:53 Crystal Ball AI Lending
01:09:11 Advice And Wrap Up
01:13:24 Hawk AI: Realtime Fraud Monitoring (sponsor)
01:14:10 Disclaimer
Comments
Want to join the conversation?
Loading comments...