The Future of Lending Is AI-Native: A Conversation with Casca CEO Lukas Haffer

By John San Filippo

Traditional loan origination systems (LOSes) have long been a source of frustration for financial institutions. Plagued by antiquated interfaces, manual processes, and a general lack of user-friendliness, they often hinder rather than help the lending process. Enter Casca, a company claiming to offer the first artificial intelligence (AI)-native LOS, designed especially for Small Business Administration (SBA) and other commercial lending. To understand what this means for the future of lending, Finopotamus spoke with Lukas Haffer, CEO and co-founder of San Francisco-based Casca.

Revolutionizing a Stagnant System 

Lukas Haffer

Haffer doesn’t mince words when describing the current state of LOS technology. “First of all, everyone I know hates their current LOS,” he stated. He attributes this widespread dissatisfaction to the fact that “all of them are ancient and horrible.” The core problem, as he sees it, lies in the heavily manual nature of existing systems. “Everything a lender, underwriter, credit officer, or closer does in an LOS today is manual,” Haffer explained, adding that even automated processes are driven by “very simplistic rules,” which are often not fully configured due to their complexity.

This reliance on manual tasks translates into significant inefficiencies. Lenders spend their days mired in emails, exchanging checklists to gather necessary documents from customers to close loans, said Haffer. In 2025, he finds this approach unacceptable, comparing it to filling out IRS forms manually instead of using software like TurboTax.

Casca’s AI-native approach aims to dismantle these inefficiencies, Haffer said, noting how Casca digitizes traditional forms, transforming them into “beautiful online forms.” Furthermore, for every document gathered, Casca “analyze[s] the entire document and automatically grab[s] the information that you really need for a loan decision.” This includes tasks like financial spreading and validating information such as business address.

Many previously manual tasks are replaced with “third-party data integrations” to confirm information, he said. The document collection process is streamlined through a digital portal, and automated follow-up messages are drafted and sent to borrowers. According to Haffer, this radically simpler and faster process significantly reduces the time needed to collect information and prepare loan decisions.

The impact on loan closing times is substantial. What used to be a 90-day average to close an SBA loan can now be condensed significantly, he told Finopotamus. “I think our average is around 10 days, and we’re going for three days,” he added. This drastic reduction in turnaround time offers a “huge service to the small businesses in America that are looking for funding, and for whom 90 days often is just not an option.”

Beyond benefiting small businesses, Casca offers a significant competitive advantage to financial institutions. Haffer contended that if a bank or credit union isn’t using Casca and still takes 90 days to close a loan, they “essentially ... no longer can exist in the lending market” when faced with competitors that can close the same loan in three days.

Beyond SBA: Expanding Commercial Lending Horizons

While Casca began with SBA lending, its capabilities extend far beyond this application. Haffer clarified, “We do anything from originating a $10,000 small business loan... to a more complicated SBA loan that might take 10 days, for example, and that goes up to $5 million in SBA funding.” The system is also designed to handle equipment loans up to $350,000, which can be fully automated, allowing borrowers to receive their money in minutes. Casca can also service commercial real estate loans up to $50 million.

For credit unions considering venturing into commercial lending, Haffer strongly recommends exploring modern lending platforms over legacy systems. He believes that a partnership with Casca, for example, would put a credit union “ahead of the game right out of the gate.” Casca can guide employees through the lending process, allowing credit unions to clearly define their credit policies, he added. Even with a smaller team, a credit union using Casca can effectively compete with larger institutions because of the system’s advanced, AI-driven automation.

Efficiency and Conversion: Tangible Results

Haffer shared a compelling example from one bank, where Casca’s faster and easier process, coupled with quicker customer follow-ups, dramatically improved conversion rates. “We saw their conversion rates of people that submit a full application go from only 8% previously to 81% with us,” he stated. This remarkable increase highlights a fundamental truth: “If someone comes to you looking for a loan, and you make it really easy for them to submit all of the information that you need, they will actually do it. If you make it really hard and confusing, the majority of them will fail and give up,” Haffer explained.

This also translates to increased efficiency in the back office. Additional, anecdotal evidence gleaned from from lenders and underwriters suggests substantial efficiency gains, with some reporting 90% of their time saved, and others saving 20 hours a week, he noted.

AI as an Enabler, Not a Replacement for Human Expertise

While Casca leverages AI heavily, Haffer emphasized that the goal is not to eliminate human involvement in credit decisions. He is a strong advocate for “making better, faster decisions, not making them worse and automatic.” He is particularly critical of over-reliance on simplistic scoring models: “This whole ‘let’s just take a FICO and an SPSS score and roll with it’ is a disaster – not good underwriting and not prudent lending.” Instead, Casca focuses on automating the analytical aspects, such as calculating the debt service coverage ratio, a strong predictor of loan repayment This automation makes it “easier for an underwriter to make a fast and good credit decision so they can review the financials all in one page very quickly.”

Ultimately, Haffer envisions Casca transforming the daily life of lenders. Instead of logging in to an empty loan file and manually sifting through emails, a lender using Casca on a Monday morning would see “a completed loan file, because the entire process has already taken place over the weekend.” Their primary task would then be to “validate that the recommended credit decision per your own credit policy is correct, and you approve the loan and sign the agreement.” This shift, Haffer concluded, is about empowering lenders to close more loans with “90% less manual effort” and experience a “way more satisfying day.”