AI Fraud Detection for Cross-Border Banking

99.97% accuracy · <50ms response · ISO20022 native

99.97%Accuracy
<50msResponse
$42BAnnual Fraud Loss
ISO20022Native

The Problem

Banks lose $42B every year to fraud.

    The Solution — Phoenix DNA™

    A 3-layer stack purpose-built for ISO20022-era banking.

    Architecture

    Ingestion

    ISO20022 native parser — pacs.008, pain.001, camt.054 at source.

    Graph ML

    Behavioral graph — every transaction is a node + edge.

    Decision Engine

    Real-time scoring · fully explainable · regulator-ready.

    Compliance

    SOC2 Type II planned · ISO-27001 · GDPR · NIS2.

    Market

    $90.07BFraud Detection Platform (2030)
    $200B+Cross-Border Payments Loss
    $308.6BInsurance Fraud (Global)

    Roadmap

    1. Q3 2026 — Pilot deployment with 2 tier-1 EU banks
    2. Q4 2026 — SOC2 Type II audit · ISO20022 certification
    3. Q1 2027 — Seed round · APAC expansion
    4. Q3 2027 — Quantum-safe key rotation in production

    Founder & CEO

    YC

    Yehuda Chen

    Founder & CEO, Phoenix Global Group

    "We're not building a firewall. We're building the immune system for global banking."

    Use of Funds — Pre-Seed $1M

    • 40% Engineering (Graph ML, ISO20022 parser)
    • 25% Compliance & Certifications (SOC2, ISO-27001)
    • 20% Pilot Deployments (EU + IL banks)
    • 15% Operations & Legal

    Business Model

    SaaS licensing per transaction volume + annual platform fee. Tier-1 banks: $250K–$1M ARR per deployment.

    Competitive Edge

    Legacy

    Rule-based · 73% miss rate · 95%+ false positives · ISO20022 retrofitted

    Phoenix DNA™

    Graph ML · 99.97% accuracy · <0.2% false positives · ISO20022 native

    Traction

    Pre-Seed · Raising $1M

    Active conversations with tier-1 banks in EU and IL.

    Security & Compliance

    ISO 27001SOC 2 Type IIGDPRNIS2CCPAPCI-DSS

    Use Cases

    Cross-Border Payments

    ISO20022 pacs.008 real-time screening.

    Card Fraud

    Behavioral anomaly detection at swipe.

    AML / KYC

    Graph-based entity resolution.

    Insurance Claims

    Pattern detection across claim networks.

    Press & Recognition

    Coming soon — pilot announcements Q4 2026.

    FAQ

    Is Phoenix DNA™ production-ready?

    Beta with pilot banks Q3 2026. GA Q1 2027.

    How is it different from legacy systems?

    Native ISO20022 parsing + graph ML, not retrofitted rules.

    What's the integration time?

    4–8 weeks for tier-1 banks via REST API + Kafka stream.

    Where is data stored?

    EU (Frankfurt/Zurich) or customer-tenant. Zero data leaves jurisdiction.

    Stay Informed

    Contact