Next-Generation Fraud Investigation Copilot

Mule Account Intelligence
Powered by GNNs

Sentinel AI maps financial ledgers as graph neural networks and evaluates transactional subgraphs in real time. Detect money mule structures and instantly write SAR compliance narratives.

0K+

Suspect Mapped

0M+

Processed API Requests

0.8%

System Accuracy

0+

Protected Corridors

sentinel-dossier-dashboard-preview.sh
romance scam alert
GNN SCORE94%
Victim
ZELLE
Mule
WIRE
Out

[ feature suite ]

Forensic Intelligence Tools

Sentinel AI combines graph neural architectures with generative narration APIs to deliver comprehensive money mule network mitigation tools.

GNN Network Tracing

Map complex multi-hop transfer chains to trace romance scams and layerings across institutional accounts in real time.

GraphSAGE Anomaly Scoring

Our machine learning pipeline uses graph neural network node classification to evaluate transaction risk topology at scale.

Gemini Forensic Explainer

Convert abstract transactional nodes and timing vectors into clean, structured human-readable case audit logs instantly.

Auto-Drafted SAR Filings

Instantly draft FinCEN Form 111 compliance narratives aligned with regulatory standards, reducing investigation times by 95%.

Network Freeze Protocol

Execute isolation directives to freeze target mule hubs and notify connected institutions with a single click.

Active Risk Diagnostics

Run automated audits over velocity anomalies, device collisions, IP deviations, and historical cybercrime tip indexes.

[ process pipeline ]

End-to-End Forensic Path

Understand how Sentinel traces suspicious multi-hop trails and dispatch compliance locks.

01

Ingestion & Parse

Financial ledger events and cybercrime notices are parsed dynamically.

02

GNN Evaluation

GraphSAGE GNN evaluates multi-hop transaction topologies.

03

Risk Telemetry

Diagnostics checklist scans for device collision and velocity spikes.

04

Gemini Synthesis

Forensic LLM orchestrates risk explanations and evidence reviews.

05

Isolation Lock

Suspect hub accounts are locked, isolating adjacent transfers.

06

SAR Submission

Auto-drafted SAR reports are dispatched to the FinCEN gateway.

[ architecture layer ]

Technology Stack & Integrations

React 19Next.js 16 (Turbopack)TypeScriptTailwind CSS v4Google Gen AI SDKFramer MotionVis.js NetworksFastAPI CorePython MLDocker ContainersGCP Vertex AIGCP BigQuery

[ user validations ]

Compliance Investigator Reviews

★★★★★

"Sentinel's GNN trace changed the game. Tracing multi-hop networks that used to take hours of manual Excel tracing now takes only seconds. The Gemini SAR generation is incredibly accurate."

Marcus Vance

Head of AML Investigations, FinTech Bank

★★★★★

"Applying GraphSAGE classification straight over transactional network subgraphs is the mathematically correct way to solve mule account layering. Sentinel's visual interface is stellar."

Dr. Clara Sterling

Lead AML Data Scientist

★★★★★

"The auto-generated Suspicious Activity Reports (SARs) are directly compatible with our FinCEN filing templates, allowing us to freeze hubs and dispatch reports with minimal friction."

Ethan Thorne

Regulatory Compliance Director

[ q&a ]

Frequently Asked Questions