Our latest articles and announcements

CompanyAugust 2026 · 6 min read

Kita raises $4.5 million to help lenders underwrite the borrowers credit bureaus miss

Kita powers AI credit assessment for some of the world’s largest microlenders, banks, fintechs, SME lenders. We raised $4.5M in seed funding led by BoxGroup to automate loan origination and underwriting analysis to unearth creditworthy borrowers that traditional systems overlook.

Read post

Kita $4.5M Seed Round, led by BoxGroup with participation from Y Combinator, Kaya Founders, US News, BEENEXT and Golden Gate Ventures

GuidesAugust 2026

Best Loan Origination Systems for Lenders in Vietnam

Much of Vietnamese origination still runs by hand: staff read every page, reconcile the figures across documents, and write every recommendation while the software records the outcome. A 2026 comparison of eight origination platforms for Vietnamese lenders, what each actually automates, and which type of institution it fits.

Kita Team10 min read

GuidesJuly 2026

Best Loan Origination Systems for CDFIs and Credit Unions

Most loan origination systems digitize the workflow while the work itself stays manual: staff still read every statement, spread every financial, and write every memo. A 2026 comparison of six origination platforms for community lenders, what each actually automates, and which type of institution it fits.

Kita Team10 min read

GuidesJuly 2026

Best Loan Origination Systems for Lenders in the Philippines

Most Philippine lenders still originate by hand: staff read every page, encode every figure, and write every recommendation while the software records the outcome. A 2026 comparison of six origination platforms for Philippine lenders, what each actually automates, and which type of institution it fits.

Kita Team10 min read

TechnicalJuly 2026

The Untapped Data Source: +7.0 Gini From Borrower Documents in Microloan Underwriting

MSME and micro lenders already collect bank statements and payslips; almost none score them. We backtested 8,000 microloans across a globally diverse pool of lenders, and borrower-submitted document signals added up to +7.0 Gini over the bureau score.

Kita Team5 min read

Community LendingJune 2026

How CDFIs Can Get Started With AI

AI adoption is a ladder, not a leap. A practical five-rung guide for community lenders — from organizing data and everyday productivity tools to document automation, AI-assisted underwriting, and predictive credit models.

Kita Team9 min read

TechnicalJune 2026

Bank Statement Extraction, Benchmarked: 9 AI Systems on Real Lending Documents

Computer vision AI models can extract every line of a bank statement almost flawlessly but still get the totals wrong. For lenders, these totals are often what the entire credit decision is built on. We benchmarked nine models against real statements and scored the signals lenders actually underwrite on.

Kita Team9 min read

GuidesJune 2026

Best Document AI Platforms for Banks & Lenders in Southeast Asia

Lending in Southeast Asia runs on messy, real-world documents — GCash screenshots, photographed payslips, bank statements in Bahasa Indonesia — that legacy OCR was never built to read. A 2026 comparison of the document AI platforms that turn that mess into decision-ready, fraud-checked data.

Kita Team11 min read

Case studiesMay 2026

How TRBank, Inc. (A Rural Bank) transformed paper loan records into structured credit intelligence

5,100 loan records. 194 fields per loan. Automated validation across the full corpus.

Kita Team12 min read

TechnicalMay 2026

Reconstructing Financial Evidence from Degraded Documents

Most OCR systems are optimized for clean enterprise documents. Real-world underwriting depends on photographed receipts, compressed bank statements, and faded invoices. How recognition-guided diffusion models recover financial evidence that traditional pipelines cannot.

Kita Team14 min read

Community LendingMay 2026

Why CDFI Underwriting Is Still Incredibly Manual

Most lending infrastructure assumes structured documents. CDFIs and credit unions deal with the opposite — fragmented records, inconsistent formats, and borrowers that legacy systems were never built to serve. Here is why that matters and what is changing.

Kita Team12 min read