Shivi Sharma spent a decade working in credit score danger at locations like American Specific and Varo Financial institution.
Sooner or later, she realized groups had been spending equal quantities of time analyzing all sorts of loans — no matter whether or not it was price $100,000 or $5 million — that means assessing smaller loans was in the end an unprofitable and time-consuming course of for lenders.
She and her husband, Utsav Shah, realized there was a chance right here.
“She watched as the overwhelming majority of small enterprise homeowners couldn’t entry the capital they wanted to develop, just because the economics didn’t work for banks,” Shah instructed TechCrunch.
“Between our abilities in constructing AI-powered decision-making techniques at scale and our experience in credit score danger and fraud danger assessments in banking in monetary companies, we realized we might apply next-gen AI agent workflows to resolve this decades-old downside,” he continued.
The married couple determined to launch Kaaj in 2024, an organization that helps automate credit score danger evaluation in order that underwriting now not takes days, however minutes. Kaaj stated it’s processed greater than $5 billion price of mortgage functions, with shoppers together with Amur Tools Finance and Fundr. The corporate introduced on Wednesday a $3.8 million seed spherical from Kindred Ventures and Higher Tomorrow Ventures.
The product works like this: A small enterprise applies for a mortgage, submitting all of the wanted paperwork (like monetary statements, financial institution statements, and tax returns) — sometimes, when this occurs, underwriters spend days manually verifying all this info and logging it into their Mortgage Origination System (LOS).
Techcrunch occasion
San Francisco
|
October 13-15, 2026
Kaaj makes use of AI to determine, classify, confirm, and arrange info into LOS. It additionally runs assessments to verify for doc tampering for the underwriter fraud crew. It integrates into current Buyer Relationship Administration (CRM) techniques like Salesforce, HubSpot, or Microsoft and even reveals a lender if a enterprise is assembly the factors of a lender’s coverage.
“This permits a crew processing 500 functions month-to-month to deal with 20,000 functions with the identical employees, making smaller loans economically viable,” Shah, the corporate’s CEO, stated.
The hope is that extra small companies will be capable to entry loans from banks as a result of it turns into extra cost-efficient for a financial institution to examine them.
Others out there embody Middesk, Ocrolus, and MoneyThumb. Sharma hopes that Kaaj will stand out from the competitors by automating the complete credit score evaluation course of somewhat than elements of it.
“We do that by deploying agentic AI workflows that mimic their groups, to assist lenders analyze end-to-end mortgage packages,” she stated.
The recent capital will probably be used to assist speed up product improvement and develop throughout unbiased and small enterprise lenders. “We’re centered on enhancing our AI agent capabilities, increasing our module choices, and scaling our buyer base of lenders and brokers past our present footprint.”
Total, Shah and Sharma hope Kaaj can in some methods “revolutionize” small enterprise lending, bringing automation to what’s nonetheless a really paper-heavy course of.
“By automating the science of credit score evaluation, we liberate human underwriters to deal with the artwork of deal-making and subjective evaluation, which is their true aggressive benefit,” he stated.
