Fintech Meetup 2026: The New Science of Credit Intelligence and Alternative Data

By John San Filippo

At Fintech Meetup in March, Finopotamus met with industry leaders in the credit data and analytics space to discuss how modern infrastructure and behavioral insights are transforming the lending landscape for credit unions. Presented herein are expert insights from these sessions.

Modernizing a 1998 Infrastructure

The underlying technology for credit reporting in the United States was largely designed in 1998 to solve problems relevant to that era, such as reporting personal loans and credit cards, explained Christian Widhalm, CEO of the New York City-based Bloom Credit. This legacy, he added, infrastructure was never intended to handle the influx of new data types – such as buy now, pay later (BNPL), credit builder products, and granular bank transaction data – that have emerged over the last decade.

Christian Widhalm

Bloom Credit bills itself as providing a modern, agnostic platform that enables bi-directional data transmission between financial institutions and the three major credit bureaus. Unlike traditional systems built around 30-day statement cycles, this platform can operate on a real-time basis. Widhalm noted that this is critical for consumers who have cured a delinquency and need their credit score to reflect that change immediately when applying for new credit, rather than waiting 30 to 45 days for a manual update.

The ‘Carrot’ Approach to Credit Building

One of the most significant opportunities for credit unions lies in reporting “alternative” trade lines that have historically been used only as a “stick” for collections, said Widhalm. He observed that utility and cell phone companies typically only report to bureaus when a consumer goes severely delinquent. Bloom Credit flips this model by allowing members to connect their checking accounts to report on-time payments for rent, utilities, and cell phone bills as positive trade lines.

This feature addresses the “chicken or the egg” problem for the 165 million Americans with limited or thin credit files. For credit unions, this isn’t just a wellness tool; it is a primary driver for member acquisition. Widhalm stated that the ability to build credit through normal daily payments is the number one feature millennials and Gen Z members look for in a checking account. This engagement translates into tangible results; institutions using the platform have seen an 11% increase in deposits and a 30% increase in credit applications within the first three months of implementation.

Mining “Behavioral Nuggets” for Custom Risk Models

While Bloom Credit focuses on the infrastructure of data transmission, New York City-based Carrington Labs specializes in the analytics used to interpret that data. CEO Jamie Twiss argued that generic credit risk models often underperform because they fail to account for the unique customer experiences and product sets of individual credit unions.

Jamie Twiss

Carrington Labs, Twiss explained, builds custom risk models by combining traditional credit files with deep transaction-level data. He highlighted the importance of finding “behavioral nuggets” within a member’s cash flow. For example, Carrington’s models look for signs of financial stress long before a missed payment occurs, such as when a member begins to compress their discretionary spending in anticipation of a tight cash month. “That has significant credit implications,” Twiss noted, adding that identifying these patterns allows lenders to intervene or adjust strategies much earlier in the cycle.

The Math of Limit Setting and Lifetime Value

A distinct challenge in lending is determining the appropriate credit limit for an approved applicant, Twiss noted. Determining whether to offer a $5,000 or $15,000 limit is a rigorous analytic problem that most lenders currently solve with generic logic, he said. Carrington Labs uses probability of default elasticity to find the “sweet spot” for each member.

“We generally find that most lenders have been under-lending to a lot of their good customers,” Twiss said. By analyzing how similar members performed with different limits and terms in the past, the platform can predict the expected customer lifetime value for a specific offer. This data-driven approach allows credit unions to safely increase their lending volume without a proportional increase in risk, he explained.

Beyond Origination: Cash Flow Servicing

Both Widhalm and Twiss emphasized that credit intelligence should not stop at the moment of loan origination. Twiss described a process called “cash flow servicing,” where institutions monitor existing books to identify members who may be in the market for a home loan or those who require a proactive outreach due to emerging stress. Similarly, Widhalm noted that “scouring” checking account data allows credit unions to provide prescriptive marketing insights – such as identifying a member who has started shopping at baby stores – to offer relevant products at the exact moment of need.

By leveraging these modern tools, credit unions can move beyond traditional credit scores to a more holistic, real-time understanding of their members’ financial lives, these experts noted.