In the world of open banking, the foundational business model – consented API access to consumer bank accounts for third parties – continues to crumble under the dual pressures of commoditization and the “pay-to-play” economics between banks and providers. Once considered an innovative technology, today, accessing consumer bank data through a secure API, and not through screen scraping, is little more than low-value raw data plumbing with little to no differentiation between providers. And the original business model did not account for the cost center of having to pay banks for the API calls – it was assumed that regulators would align with the notion that the consumer owns its banking data and the consumer could share the same (share for free) with whomever it wanted. For many companies in the space, these two dynamics have created a new reality – open banking is a tool, not a monetizable, value added product.
The consequence of this is real (see more here). There’s been significant consolidation. In 2025 alone, Visa sold off its open banking unit, Evestnet sold off Yodlee, and Mollie acquired GoCardless, though not because it wanted or needed its API access, but rather its Account-to-Account payments functionality. In sum, this plays to a broader market thesis for the entire open banking industry – it’s incumbent on open banking companies to recognize the tool vs. product dilemma, and learn to use the tool to create new products. For the industry to remain viable, this will be a critical pivot, and that’s exactly what one industry leading open banking and data aggregation company has done.
Plaid. A winning paradigm.
This week Plaid showcased its Fall Product Release, which included, among new AI models for payments and fraud detection, an updated model for credit scoring – LendScore2 – for fintechs and smaller financial institutions that want to compete in the consumer financing segment, including auto loans, HELOCs, mortgages, and banking and credit products for “no-file” and “thin file” applicants (many of these underbanked are gig workers and immigrants).
This second generation LendScore product boasts 42% more predictive power.
Note: On “predictive power”, what was measured and how this was calculated is not entirely clear, however it is assumed in part this was done using a KS (Kolmogorov-Smirnov) score which Plaid used to evaluate LendScore2’s predecessor. A KS score is a statistical metric that measures how well a credit risk or scoring model separates “good” borrowers from “bad” borrowers (those who defaulted).
The investment and the product launch show that Plaid truly grasps the “tool vs. product” conundrum in open banking, and this big bet on new and enhanced product is exactly the aggressive innovation stakeholders should expect from the dominant open banking player in the US market (measured by bank, consumer and fintech connectivity). Plaid is doing precisely what I alluded to above: using the data it has access to to create value added products.
Cash flow underwriting.
At its core, LendScore2 powers cash flow underwriting, a different approach to underwriting consumer loan products and a proven high-value signal. Traditional underwriting relies heavily on data from the three major lender reporting agencies – Equifax, Transunion, and Experian, and the resulting FICO score derived from that data by Fair Isaac Corporation. Taken together, this amounts to a “lookback” methodology that reflects a borrower’s willingness to pay. This is sufficient for consumers with a long borrowing history.
But what about consumers who need financing who don’t have a borrowing history, the no-file and thin-file consumers I mentioned above? Cash flow underwriting is a solve for this constituency. It’s essentially real-time auditing of a consumer’s bank account, extracting signals from actual cash flows, from when bills are paid to who a consumer is paying.
Note: the “who” a consumer is paying speaks to responsibility and risk, is the consumer paying DraftKings or CVS?
Additionally, under a cash flow underwriting paradigm, constant monitoring also calculates a running payments-in (income) vs. payments-out (expense) analysis to determine whether there’s a positive balance that can support a new loan obligation. Cash flow underwriting shows lenders and originators a consumer’s ability to pay back a loan, rather than its willingness.
The real takeaway.
The power of cash flow underwriting and Plaid’s LendScore2 credit scoring notwithstanding, the real story, or at least the real commercial story, is that Plaid has created a product that it can sell to a non-bank consumer loan origination market (fintechs especially) that seems to continuously expand (via Transunion, see here), Plaid has successfully navigated the strategic pivot from tool to product, solving both the commoditization and differentiation headwinds to the open banking industry writ large.
The API vs. screen scraping debate is over. This is the future of open banking.
Plaid, through its innovative products and strategic vision, has become the preeminent open banking player in consumer facing segment and it deserves a lot of credit (pun 100% intended) for its market leadership, service to the underbanked, and its evolution to what can best be described as an alternative “alt” credit bureau. Through its value added product development, it has created the roadmap to viability for the entire open banking industry.
