Lenders already know the value of cash flow data, and increasingly, they’re recognizing how a network view of financial activity can reveal deeper patterns in behavior and risk. But accessing those insights has historically required borrowers to connect a bank account at the point of application, which can contribute to borrower drop-off.
For lenders to use cash flow and network-level insights at scale, we focused on two things: getting permissioned data and insights without extra steps for the borrower, and turning them into a precise, decision-ready view of risk.
Here's what's new:
1. Instant Link: A more seamless way for borrowers to securely permission their cash flow and network insights for future credit applications. Lenders can then access those insights in under two seconds.
2. LendScore 2: A next-generation family of credit risk scores. It includes Ls2, our upgraded core model that delivers 42% greater predictive power than traditional credit data alone, specialized models for Auto, Short Term, and Home Lending, and LendScore Arc—our first transformer-based credit risk model.
Instant Link: A faster, secure way to access permissioned cash flow and network insights
Today, consumers expect digital financial experiences to be fast, secure, and easy. And for 1 in 2 US adults with a bank account, Plaid is already a familiar way to connect their financial accounts.
Traditional credit data can be accessed with a soft pull in seconds. Cash flow, on the other hand, has typically required a borrower to connect their financial account during the application. To bring cash flow into more credit decisions, it needs to be just as readily available.
That’s what Instant Link makes possible. Consumers can now consent to lenders accessing their cash flow data through Plaid’s Consumer Reporting Agency (CRA), so they no longer have to reconnect their bank account on subsequent credit applications. When they return, all lenders need is basic borrower information—like name, address, and phone number. Plaid matches the borrower to their network profile, and returns permissioned insights back to the lender in under two seconds.
Consumers have already started permissioning their cash flow data through Plaid’s CRA. As more people connect their accounts with Plaid, lenders can deliver a more seamless borrower experience while putting cash flow and network insights to work across the lending lifecycle—not just for second-look underwriting, but from prequalification through servicing.
A next-generation family of credit risk scores
Last year, we introduced LendScore: our first-of-its-kind credit risk score built on cash flow and Plaid Network data. Designed to be used broadly across consumer lending, our first version of LendScore delivers an average 25% lift in predictive performance when used alongside traditional credit data, raising the bar for what lenders can achieve using cash flow and network insights.
The next generation of LendScore introduces an updated 1,000–2,000 score range, providing greater precision to differentiate risk across borrowers. The expanded family of credit risk models starts with Ls2, our upgraded core model, which delivers 42% greater predictive power than traditional credit data alone.
For Auto, Short Term, and Home Lending, we introduced specialized models optimized for the distinct risk dynamics of each market. We built specialized models only where a more tailored approach delivered meaningful incremental performance gains. For example, in testing our model among deep-subprime auto applicants, we saw that Ls2 Auto can lower delinquency by 26% compared with traditional credit data alone at the same approval rate.
In the mortgage and HELOC market, Ls2 Home Lending can approve 6.3% more borrowers than traditional credit data alone at the same level of risk. And for shorter-term products like BNPL, Ls2 Short Term uses a 60+ DPD within a six month performance window and can approve 63% more borrowers than traditional credit data alone at the same level of risk.
For lenders ready to put more advanced risk modeling to work, we’re introducing LendScore Arc, our first transformer-based credit risk model. Arc uses Plaid’s Sequential Foundation Model to take a different approach from the rest of the LendScore family. It learns by capturing additional signals from the order, timing, context, and interactions between transactions, then combines them with the established underwriting features already used by LendScore to assess credit risk.
The result is our best-performing credit risk model yet. In early testing, Arc showed especially strong performance at opposite ends of the credit spectrum, delivering 20% predictive lift over our core model for deep-subprime borrowers and 24% lift for super-prime borrowers.
That performance comes with greater model complexity, which is why we’ve invested in a cutting-edge explainability framework using Integrated Gradients—helping lenders understand the financial factors contributing to each prediction and translating them into clear, standardized reason codes. We pair that explainability with rigorous fair-lending testing and controls in partnership with Fairplay, to support responsible use of our models at scale.
As with the broader LendScore family, we work with lenders to choose the model that best fits their performance goals and governance requirements. Every LendScore is delivered through Plaid’s CRA, with consumers retaining applicable rights under the FCRA.
The next phase of cash flow underwriting
Instant Link and the next generation of LendScore are powered by the same underlying advantage: the breadth and depth of the Plaid Network. As more consumers choose to permission their financial data through Plaid, we can make access to insights faster for the specific consumer and lender involved, while also continuing to build more powerful risk insights on top of it.
Together, these advances make it possible to put cash flow and network insights to work in more lending decisions, without adding another step for borrowers.
Get started
See how Instant Link and the LendScore 2 can help you bring cash flow data into more lending decisions and better predict risk across the loan lifecycle. Get in touch with our team.