How Vantrace works

Vantrace integrates into your existing systems to monitor transactions, detect risk, and generate regulatory reports in real time. No system replacement required.

Connect → Ingest → Score → Alert → Report
REST API integration
No system replacement
2 to 4 week deployment
Implementation roadmap on request
Engine 01
Fraud Detection

Real-time scoring on every transaction using ensemble ML models trained for Nigerian transaction patterns. Every decision is explainable and auditable.

XGBoost + Isolation Forest Ensemble approach combining supervised classification with unsupervised anomaly detection for higher accuracy on novel fraud patterns.
Sub-200ms scoring latency Transaction scoring completes before authorization, enabling real-time blocking without adding friction to genuine customers.
Mule network detection Graph-based analysis identifies coordinated mule account networks, not just individual suspicious transactions.
SHAP explainability Every alert includes a SHAP explanation showing which features drove the decision, required for regulatory review and audit.
False positive rate below 3% Threshold tuning and feedback loops reduce alert fatigue for your compliance team.

Capabilities

Real-time scoring Ensemble ML Mule detection Graph analysis SHAP explainability Alert management
Scoring latency <200ms
False positive target <3%
Models XGBoost + IF
Engine 02
AML Compliance

Automated AML monitoring from transaction-level detection through to case management and STR generation. Covers the full CBN AML/CFT workflow.

Structuring detection Identifies structured transactions designed to evade CTR thresholds, including split deposits, layered transfers, and smurfing patterns.
PEP and sanctions screening Screens customers against OpenSanctions, UN, OFAC, and NFIU watchlists. Continuous monitoring on existing customers, not just onboarding.
CDD and EDD risk profiling Dynamic customer risk scoring that escalates to Enhanced Due Diligence when behaviour or profile changes warrant it.
Automated STR generation Suspicious Transaction Reports are drafted automatically from case data, reducing manual writing time from hours to minutes.
Threshold-based alert rules Configurable rules engine aligned to CBN typologies. Rules can be updated without re-deployment.

Capabilities

Structuring detection PEP screening Sanctions (OpenSanctions) CDD profiling EDD escalation STR drafting Alert rules engine
Watchlist sources OFAC, UN, NFIU
Screening type Continuous
STR generation Automated
Engine 03
Regulatory Reporting

Generates, formats, and tracks all regulatory reports required for NFIU and goAML submission. Every report is tied to an auditable case and retained for 5 years.

STR, CTR, and SAR in PDF and JSON Reports generated in the formats required by NFIU. PDF for human review, JSON for machine processing and submission.
goAML XML format Native goAML XML generation for direct submission to the NFIU goAML portal, removing manual data entry from the filing process.
Submission status tracking End-to-end tracking from case creation through to filed status, with timestamped audit entries at each stage.
5-year audit trail All decisions, reports, and case actions are retained for the full 5-year period required under CBN and NFIU regulations.

Report formats

STR CTR SAR goAML XML PDF JSON
Retention period 5 years
Submission tracking End-to-end
Audit trail Timestamped

Infrastructure

Built to meet Nigerian data residency requirements and NDPA 2023 obligations from day one.

Deployment

  • AWS or GCP deployment
  • 99.9% uptime SLA
  • Data residency in Nigeria
  • Encrypted APIs (TLS 1.3)

Compliance and Privacy

  • NDPA 2023 compliant
  • 5-year audit log retention
  • Role-based access controls
  • No data shared with third parties

See it running on your transaction data

Request a demo and we'll walk through each engine with your specific use case.

Request a Demo info@vantrace.io