Kita is an AI-native underwriting platform that automates document intake, extraction, verification, and credit assessment for lenders serving emerging markets and underserved borrowers. Using vision‑language models, it captures data from any document type, detects fraud, and generates policy‑aligned credit drafts while allowing lenders to retain final decision authority.
Funding
Funding not disclosed
Founders
Product
Problem
Borrowers in emerging markets often lack digital financial records, resulting in fragmented, handwritten, or photo-based documents that traditional underwriting systems cannot process efficiently. Lenders must manually reconcile and verify these documents, leading to slow loan approvals, high labor costs, and increased risk of fraud.
Solution
Kita provides an AI-native underwriting platform that automates the entire document intake, extraction, verification, and credit assessment workflow. Borrowers can submit documents via WhatsApp, SMS, email, or web portals, and the AI Credit Officer interacts in the borrower’s language to ensure complete files. The Capture module uses vision‑language models to read over 50 document types, extract structured data, and flag tampering without relying on template‑based OCR. Verified data and fraud signals feed into the AI Underwriter, which generates a draft credit assessment calibrated to the lender’s policy and cites source documents. Lenders retain final decision authority while benefiting from faster, more accurate, and scalable underwriting.
Target Audience
Primary customers are banks, fintechs, CDFIs, and other credit providers serving SME, consumer, and commercial borrowers in emerging markets such as the Philippines, Indonesia, Mexico, South Africa, and the United States.
Features
- AI Credit Officer handles multilingual borrower communication and document completeness checks across WhatsApp, SMS, Viber, and email.
- Capture engine extracts structured fields, confidence scores, and tamper signals from any document format using vision‑language models, supporting 50+ market‑specific document types.
- Cross‑document fraud detection validates IDs, tax filings, and statements, identifying inconsistencies and tampering.
- AI Underwriter produces policy‑aligned credit assessments with evidence citations, returning results in JSON for easy integration with LOS/CRM systems.
- Single credit‑balance pricing model covers all modules, eliminating per‑API surcharges and seat‑based licensing.
- Forward‑deployed engineering team integrates the platform into existing workflows and provides support until production stability is achieved.