Kita
Category: AI in Fintech
AI-native credit assessment platform that turns messy borrower financial documents into fraud-checked, decision-ready signals for lenders in emerging markets. Kita was founded in 2025. The company is led by Carmel Limcaoco. Based in San Francisco, USA. Team size: 1-10. Total funding raised: $4.5M. Latest round: Seed. Key investors include BoxGroup, Y Combinator, Golden Gate Ventures, BEENEXT, US News Digital Ventures, Kaya Founders, Genting Ventures, Apex Star Capital.
- Founded
- 2025
- Headquarters
- San Francisco, USA
- Team size
- 1-10
- Total funding
- $4.5M
Value proposition
Automates application completion, document verification, and underwriting from unstructured financial documents (e-wallet records, bank statements, payslips, photos) using vision-language models, so lenders can underwrite underbanked borrowers in minutes instead of hours.
Products and solutions
Kita Capture (vision AI document intelligence + fraud), AI Credit Officer (24/7 borrower intake agent via WhatsApp/SMS/email), AI Underwriter (drafts decision-ready credit memos), AI Document Risk Engine (predictive credit signals), Intelligent LOS (AI-native loan origination system), AI-Powered Application, AI-Powered Borrower Portal.
Unique value
Hyperlocalized AI credit assessment built for emerging markets where banking APIs don't exist — reads any document format (photos, scans, screenshots, e-wallet exports) with vision-language models, surfaces hundreds of alternative-data risk signals scored by Information Value, and keeps the human lender in the loop with every claim cited.
Target customer
Lenders, banks, fintechs, microfinance institutions, CDFIs, and community/SMB lenders operating in emerging and undertapped markets (Southeast Asia, Latin America, US).
Industries served
Financial services / lending: microfinance & small-ticket consumer, SME & commercial lending, CDFI & community lending.
Technology advantage
Vision-language models outperform legacy OCR on noisy real-world documents; a learning engine that links document-level signals to repayment outcomes for compounding risk-model improvement; hyperlocalized to markets like PH, ID, MX, ZA, US; SOC 2 Type II and ISO 27001 in progress; in-region data residency.
How they differentiate
Unlike legacy OCR that breaks on real-world inputs, Kita uses vision-language models to read any document format, extracts hundreds of alternative-data risk signals scored by Information Value, and is hyperlocalized to emerging markets where banking APIs don't exist — with a human-in-the-loop lender decisioning model.
Main competitors
Legacy OCR/document-processing tools, manual underwriting workflows, alternative AI underwriting tools (e.g., Trace, Riff.ai, ChatInDoc, Gumloop).
Key partnerships
Y Combinator (W26 batch), BoxGroup (lead investor), Golden Gate Ventures, BEENEXT, US News Digital Ventures, Kaya Founders, Genting Ventures, Apex Star Capital
Notable customers
TRBank (Philippines rural bank; 5,100+ credit folders processed)
Major milestones
Accepted into Y Combinator Winter 2026 batch, raised $4.5M seed led by BoxGroup, launched AI-native underwriting platform live across Southeast Asia, Latin America, and US.
Growth metrics
Processed >$130M loan volume; 100K+ borrower files; live in PH, ID, MX, ZA, US; 15+ new enterprise pilots (July 2026).
Market positioning
Early-stage (YC W26) AI credit-assessment infrastructure player targeting the underbanked in emerging markets, live in PH, ID, MX, ZA, and US with 100K+ borrower files processed.
Geographic focus
Southeast Asia (Philippines, Indonesia), Latin America (Mexico), United States (CDFIs/community banks), South Africa.
About Carmel Limcaoco
Ex-Apple (computational audio/music product, iOS); Stanford Symbolic Systems & Music + MS Computer Science; co-founded DAHA (Stanford's first marketplace); UN awardee at 16; launched first Product Fellowship in the Philippines.
Latest news about Kita
More AI in Fintech companies
Official website: https://www.kita.ai/
