Skip to main content
Loading navigation
Back to challenges

Cross-Border AML & Sanctions Monitoring Agent with Google ADK

Foreign banking institutions expanding into North America face complex OCC sanctions and AML monitoring obligations across international payment channels. Build a cross-border wire transaction monitoring system using Google ADK and Gemini multimodal capabilities to identify structural money laundering schemes, achieving 97% detection precision.

CybersecurityHosted by Vera
Sign in free to start

What you are building

Utilize Google Agent Development Kit (ADK) and Gemini 2.0 to perform multimodal transaction audit and cross-border SWIFT wire compliance verification.

How work is evaluated

Evaluation

Evaluates SWIFT transaction parsing accuracy and trade-based money laundering anomaly detection.

Datasets

Public datasets, plus the private datasets you have access to.

How submissions are scored

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

TBML Anomaly Flagging

100% of injected trade-based money laundering anomalies must be flagged.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

detection_precision

Precision rate of cross-border AML flagging • target: 0.97 • range: 0.94-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

  • Implement Google ADK agent modules for real-time SWIFT payment parsing

  • Leverage Gemini multimodal features to review supporting trade invoice documents

  • Design automated regulatory alerting workflows compliant with OCC guidelines

Resources

Dataset notes

Dataset of 2,000 ISO 20022/SWIFT wire records paired with synthetic trade invoices containing hidden TBML patterns.

Challenge setup

Evaluates SWIFT transaction parsing accuracy and trade-based money laundering anomaly detection.

Challenge input

SWIFT message string and invoice document payload

Google ADK

Assigned agent framework for Google Gemini ecosystem.

Evaluated output

JSON audit result with TBML risk indicators and pass/flag decision

Evaluation checks
  • 100% of injected trade-based money laundering anomalies must be flagged.
  • Precision rate of cross-border AML flagging • target: 0.97 • range: 0.94-1
Configured evaluation evidence
  • Benchmark: SwiftAML-Bench
  • Detection Precision target: 0.97
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • Google · Required
  • Google ADK · Optional
  • Google Jules · Optional
Orchestration
  • Google Jules · Optional

Evaluation contract

  • TBML Anomaly Flagging · Weight 1
  • detection_precision · Weight 1

Metrics

  • detection_precision · Target 0.97

Benchmarks

  • SwiftAML-Bench
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: SWIFT message string and invoice document payload

    Output: JSON audit result with TBML risk indicators and pass/flag decision

    Execution harness: Python

Evaluation tools

  • Google ADK · Agent Development Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Cross-Border AML & Sanctions Monitoring Agent with Google ADK