How Community Institutions Are Closing the SMB Lending Gap
By Nick Holland, Contributor at The Financial Brand
Simple Subscribe
Subscribe Now!
Small businesses are the backbone of the American economy, yet they remain stubbornly underserved by the institutions that claim to champion them. Traditional underwriting treats a $50,000 loan request with the same rigor as a $10 million commercial deal, creating unit economics that simply don’t work for smaller credits.
The result: An enormous lending gap, with four out of five small businesses being declined by larger banks when they don’t fit neatly into an automated scoring box.
But a wave of community banks and credit unions, armed with new technology partners and cashflow-based underwriting models, are proving that profitable SMB lending at scale is becoming a powerful engine for primacy.
The early results are striking: Doubled approval rates, 10x throughput on the same headcount, and improved risk outcomes to boot.
Need to Know:
- Small business loans, often for small amounts, can be laborious and expensive to process, as many small business owners lack rigorous financials and full documentation.
- Smaller institutions in turn often have limited capacity and resources to devote to this line of business, choking off or turning away demand.
- Innovative lenders are bypassing the problems by mining applicants banking histories to get a direct look at cash flow and their ability to service loans. AI is lending a helping hand.
The Problem: A $50K Loan Shouldn’t Cost as Much to Process as a $10M One
The traditional SMB lending model is broken. Small business owners often lack the financial sophistication to produce the tidy financial statements that conventional underwriting demands. Meanwhile, the labor-intensive process of gathering, analyzing, and decisioning those applications makes small-dollar loans prohibitively expensive to originate.
“We were treating $50,000 loan requests the same way we were treating $10 million loan requests,” says Bill Cunningham, executive vice president of business and commercial banking at Vancity. “The quality of information, particularly as you come down market, is less certain.”
Brian Devereux, senior vice president and chief lending officer at Unitus Community Credit Union in Portland, Oregon, saw the same dysfunction. His team would spend weeks exchanging emails with applicants who didn’t know the difference between a balance sheet and an income statement. “Small businesses just don’t speak financials,” he says. “They know their trade, but they don’t have the time and gumption to do it.”
The consequences ripple outward. Alex McLeod, founder of Parlay, an AI-powered lending platform, puts it simply: “That capacity problem is rampant. At every credit union and community bank I’ve talked to, that is the limiting factor.”
The Fix: Cashflow-Based Underwriting in Minutes, Not Weeks
The breakthrough for many institutions has been a shift from historical financial statement analysis to real-time cashflow data. Instead of requesting tax returns and audited financials, lenders connect directly to an applicant’s bank accounts and transaction data, building a predictive picture of the business’s ability to service debt.
Vancity piloted this approach in 2017 with Judi.ai, a cashflow-based underwriting platform. The results were transformative. With effectively the same headcount, the credit union went from processing 40 to 60 small business loan applications per month to more than 400. “Really, a 10x increase in the number of small business members where we were actually processing loan applications,” Cunningham says.
Crucially, the risk profile didn’t deteriorate. “Not only has there not been a corresponding increase in probability of default, but as a percentage basis, it’s actually improved,” he says. Better data, it turns out, produces better decisions.
At Unitus, the transformation was equally stark. Before adopting Judi.ai, the credit union approved roughly 30 percent of SMB applications. After implementation, that rate doubled to 60 percent. The institution booked 126 loans in its first year on the platform, with decision times collapsing from weeks to as little as eight hours.
Bottom line: “All you need is 12 months of checking account data and a credit pull,” Devereux says. “It really is that straightforward.”
AI Does the Legwork. Humans Make the Decisions
A common thread across every implementation is that technology handles the drudgery while relationship bankers remain at the center of the process. AI aggregates documents, calculates financial ratios, and generates credit-ready files, but the human lender retains decision-making authority.
“Our mission is to help those relationship bankers stay relationship bankers, but just make it high tech,” McLeod says. At Parlay, the AI prepares lender-ready files so loan officers can focus on advising customers rather than sifting through paperwork. The results are already visible in the field.
Proof point: Pathway Lending, a Community Development Financial Institution in Nashville that specializes in $50,000 to $75,000 loans for underserved small businesses, launched a campaign using Parlay’s platform and received more than 30 applications in under 24 hours, with several moved to approval the same day. Borrowers connected their financial accounts and tax transcripts in real time, and the system triaged the data, calculated financial ratios, and delivered a lender-ready file with no back-and-forth document requests required.
Roger Vincent, co-founder of UK-based Bourn, frames the opportunity more bluntly. His firm has built what it calls the “always-on audit” — continuous monitoring that connects a borrower’s accounting system, bank account, and credit bureau data, replacing the expensive periodic site visits that traditional commercial banking relies on. “We’ve gone through 15, 20 years of disruption within the fintech space, but now’s the time that the banks can fight back,” he says. “Our strategy is to disrupt the disruptors. We’ve absorbed all of that learning, and we’ve got all the tools at our disposal.”
Daniel Goldstone, CEO of Rangeteller, emphasizes that transparency in AI-driven lending is non-negotiable. His platform’s machine learning models are fully explainable, a critical distinction from black-box alternatives. “If we have machine learning or AI, it’s got to be explainable,” he says. “The advanced machine learning model that we have is fully transparent.” For credit committees and regulators, that visibility matters enormously.
The Primacy Payoff: Lending is the Gateway to Full Relationships
The strategic prize extends well beyond loan originations. Institutions that crack SMB lending are finding it becomes the anchor for deep, multi-product relationships that are extraordinarily difficult for competitors to dislodge.
Vancity’s data tells the story: more than 94 percent of business members approved for loans over the past four to five years remain active members, and in most cases, active borrowers. Portfolio composition skews heavily toward lines of credit, which means borrowers also maintain active operating accounts. “When it comes to primacy, if you’ve got their day-to-day business banking, that opens the door to real deep-rooted stickiness,” Cunningham says.
At Unitus, the mission carries additional weight — 83% of 2025 bookings were to women, minority or veteran owned businesses. The credit union has made a five-year strategic commitment to lend $5 million to traditionally underserved small businesses. “It’s the foundation of the community,” Devereux says.
McLeod sees the cross-sell opportunity embedded in the lending process itself. The data gathered during loan applications — transaction history, cash flow patterns, existing account relationships — reveals additional products each borrower could benefit from. “There’s a lot of data that we’re pulling that doesn’t just have to do with the loan,” she says. “Traditionally, your average loan officer is not going to go looking for that type of data. They don’t have time. We’re just serving it up to them.”
What is the Competitive Window?
The SMB lending gap represents both an urgent problem and a finite opportunity. Fintechs and alternative lenders continue to capture market share with speed and simplicity, even if their capital costs result in higher pricing for borrowers. Goldstone estimates that smaller lenders have lost ground to the tune of roughly $50 billion in the past decade across personal loans, mortgages, and small business credit. Among Rangeteller’s clients, implementing its transparent AI framework has increased loan approvals by around 20 percent with zero additional risk on day one, even before optimizing strategies.
Steve Kietz, managing partner at Woodbury Advisors, sees AI’s biggest near-term impact not in credit decisioning alone, but in the operational layers surrounding it — reading and summarizing applications, prioritizing deal flow, and matching applications to the best-suited underwriters. The sub-700 FICO small business borrower, he notes, is still primarily served by the merchant cash advance industry, leaving room for community institutions to capture creditworthy borrowers that traditional scoring overlooks.
Bottom line: Community institutions have the cheapest capital and the deepest community relationships — what they’ve lacked is the technology to make small-dollar lending economically viable. That technology now exists. The institutions moving fastest are proving that community-based lending and modern efficiency are a single circle Venn diagram.
