AI KYB Merchant Screening
Multi-page prototype for merchant onboarding and risk screening. It operationalizes the six AI modules, cross-module controls, data-governance requirements, risk decisioning, EDD, privacy-by-design, ROPA, DPIA, audit, SLA and PoC validation.
One screening case, six AI modules, one explainable merchant-risk decision.
The prototype separates functional coverage from claims that must be measured in PoC. Accuracy, latency, source coverage and throughput are shown as validation items unless backed by an actual integrated engine.
Six Required AI Modules
Each module has a dedicated page with requirements, RFI questions, privacy controls and API schema.
OCR & Document AI
NIB, SIUP/OSS, Akta and NPWP extraction, document classification, tamper/recapture detection, cross-document consistency, field confidence and human review.
Website Intelligence
Business process, products/services, declared-vs-detected LOB, high-risk/prohibited categories, WHOIS/domain history, parking/scam indicators and Bahasa Indonesia analysis.
Google Review Sentiment
Rating/review trends, Indonesian sentiment, complaint/fraud themes, suspected review manipulation and EDD-ready negative-theme summary.
Adverse Media
Local/international negative news, entity resolution, false-positive reduction, deduplication, relevance ranking and PEP/sanctions context.
Social Media Sentiment
Legally available social sources, fraud/controversy signals, Indonesian sentiment and privacy-minimized collection.
Image Safe Search
Adult/explicit, violence, racy, medical and spoof categories with configurable thresholds and private processing options.
End-to-End Processing Flow
Illustrative architecture designed for evidence lineage and human-accountable decisions.
Enterprise investigation & lifecycle capabilities
Added to move the prototype from requirement representation toward an end-to-end KYB investigation and decisioning experience.