Home / Projects / Plus8Soft AML

Plus8Soft AML

An AI-powered anti-money laundering platform that detects suspicious transactions in real time — from open-source training pipeline to enterprise-grade compliance engine.

Discuss Your Project

Summary

Intelligent AML Detection at Scale

Money laundering costs the global economy an estimated 2–5% of GDP annually, yet legacy rule-based systems generate up to 95% false positives — burying compliance teams in noise while real threats slip through. Plus8Soft set out to build a better approach: an AML platform that combines graph-based transaction analytics with machine learning to surface genuine risk, not just volume.

The result is Plus8Soft AML — an open-core platform with a free Community Edition for model training and research, and an Enterprise Edition built for production deployment in regulated financial institutions. AML is available as an open-core platform: a free Community Edition for offline model training and research, and an Enterprise Edition for production deployment in regulated financial institutions.

Enterprise Edition

The Enterprise Edition extends the core detection engine with production infrastructure for regulated financial institutions:

• Real-time scoring API (REST / gRPC) — sub-100ms transaction evaluation
• Streaming ingestion — Kafka, Pulsar, Kinesis, native ISO 20022 support
• Regulatory reporting — automated SAR / CTR / STR generation (FinCEN, FATF, EU 6AMLD, UK MLR)
• Sanctions screening — OFAC, EU, UN, HMT watchlists
• Investigator case management UI — visual graph exploration, case workflows, decision documentation
• Model governance & MLOps — drift detection, champion/challenger, model registry
• 7-year audit logging — regulator-grade traceability for examinations
• Core banking connectors — Temenos, Finastra, Finacle, FIS, Fiserv, Mambu, SWIFT
• Enterprise security — SSO (SAML / OIDC), RBAC, multi-tenancy, air-gapped deployment
• 24/7 support — named TAM, LTS branches, IP indemnification

Industry

Fintech

Services Used

AI/ML Engineering, Platform Development, Data Engineering

98.6%

Detection Precision

777K+

Transactions Analyzed

95%

Fewer False Positives vs. Rule-Based Systems

3

Published Research Papers

"We built this because we saw the same problem at every financial institution we worked with — compliance teams drowning in false alerts while actual suspicious patterns went undetected. The goal was never just a better model. It was a better system."

Timur Iusubaliev, Plus8Soft

The Deliverable

From Research Pipeline to Production Compliance Platform

We engineered a complete AML detection system that spans the full lifecycle — from offline model training and experimentation to real-time transaction scoring and regulatory reporting. By releasing the training pipeline as open source, we made the core technology accessible to researchers and smaller institutions while building the enterprise layer for banks and fintechs that need production-grade performance, audit logging, and regulatory compliance.

Community Edition: Open-Source Training Pipeline

A Python-based ML pipeline supporting plain and graph dataset schemas. Builds directed transaction graphs using NetworkX, extracts features (in/out degrees, PageRank, cumulative totals per party), applies correlation-based filtering, and trains RandomForest, XGBoost, and LightGBM models. The best model by ROC-AUC is saved with full metrics for offline inference. Licensed under Apache 2.0.

Real-Time Scoring Engine

Enterprise-grade API (REST and gRPC) delivering sub-100ms transaction scoring. Ingests streaming data from Kafka, Pulsar, and Kinesis with native ISO 20022 support. Every transaction is evaluated against trained models in real time, enabling immediate blocking or flagging of suspicious payments before settlement.

Regulatory Reporting & Compliance Module

Automated generation of Suspicious Activity Reports (SAR), Currency Transaction Reports (CTR), and Suspicious Transaction Reports (STR) aligned with FinCEN, FATF, EU 6AMLD, and UK MLR frameworks. Integrated sanctions screening against OFAC, EU, UN, and HMT lists. Seven-year regulator-grade audit logging ensures full traceability for examinations.

Investigator Case Management UI

A purpose-built interface for compliance analysts to review flagged transactions, visualize relationship graphs, document investigation decisions, and manage case workflows. Advanced graph analytics surface complex laundering typologies — structuring, smurfing, layering, and community detection — giving investigators the context to act decisively.

Technology Stack

See How It Works

Request our free AML research paper or talk to the team about deploying Plus8Soft AML in your organization — no commitment, no sales pitch.