Engineered for AML performance evaluation
Syntheticr provides high-fidelity synthetic financial data designed specifically for testing, training, and benchmarking AML systems and models.
OVERVIEW
What the data includes
A complete testing environment
Syntheticr datasets include the full range of datasets required for testing:
Transaction data
Entity and account data
Risk intelligence data
Company registry data
Multi-institution network structures
Together, these components create a realistic environment for evaluating how AML systems behave across the different scenarios and conditions found in the real world.
How the data is designed
Precisely-engineered for AML
Syntheticr data is designed to support objective evaluation, not just realism:
Financial behaviour and transaction flows
Embedded financial crime activity
Known ground truth held internally for evaluation
Multiple customer types, entity relationships, and network structures
Typologies reflecting real-world money laundering behaviour
Realistic-looking data is not enough. The value comes from being able to test systems against embedded activity in a way production data typically cannot support.
COVERAGE
From single institutions to full ecosystems
Syntheticr provides datasets of varying scope to support different testing objectives.
Single Financial Institutions
Used for baselines, focused validation, and controlled comparisons within a single institution context.
Full Financial Ecosystem
Used for testing across institutions, counterparties, transactions, and network-level behaviours.
FEATURES
A national-scale economic simulation
Syntheticr datasets are derived from a continuous simulation spanning a time period of 24-months. They include:
A full two-year transaction history
Risk intelligence for the first 18 months
A final greenfield period for unbiased testing
Different institution sizes and operating conditions where relevant
This design supports testing and evaluation against available intelligence and beyond, to protect against over-fitting and prove transfer learning.
Test against real-world behaviours
Syntheticr is designed to test detection capabilities across network typologies and money laundering methodologies. Datasets include:
Smurfing
Mule activity
Layering
Circular flows
Multi-institution networks
This allows teams to evaluate detection of what really matters.
Built to work in existing environments
Syntheticr datasets are designed to fit into existing AML systems and model development pipelines.
Delivery options include:
Secure Download
Delta Sharing
Cloud Syncing
This allows teams to test their systems and models with minimal disruption to existing processing workflows.
CUSTOMISATION
Syntheticr supports custom variants for specific use cases and requirements
Popular customisations
Data quality degradation
Custom network typologies
Custom criminal methodologies
Temporal extensions
Specialist use cases
Technology demonstrations
Hackathons
Employee training
Volumetric/Unit Testing
Custom variants are available on request, and we are happy to discuss your needs.
GET STARTED
Start with a baseline
For most teams, the right starting point is a performance baseline.
Run a Syntheticr dataset through your current AML system or model and receive a quantified scorecard showing what your system detects, what it misses, and any performance gaps.