Defence Company
Developing Financial Crime Systems
A major defence company set out to build a financial crime detection product combining homomorphic encryption (HE) and entity resolution (ER).
The vision was ambitious: allow multiple organizations to detect illicit activity across data silos while keeping sensitive information encrypted end-to-end. To get there, the company needed realistic financial crime data—not because live data was restricted, but because finding high-quality datasets with labelled crime activity is extremely difficult.
Syntheticr Ecosystem filled that gap with synthetic financial transaction and risk-intelligence, containing labelled financial crime typologies at real-world scale and complexity. These datasets became the benchmark that guided development, validated performance, and proved whether the encrypted ER approach worked as intended.
Building HE-enabled entity resolution for financial crime is only useful if it works across a wider network of interconnected transactions, creating a series of challenges for developing and testing the solution:
Data Sourcing
The team needed realistic, high-volume financial transaction networks with labelled criminal activity, but sourcing that data was extremely difficult due to its sensitivity and the scale required.
Data Complexity
The product had to show it could link entities and uncover suspicious networks across datasets owned by different parties, so the training data needed to include multiple financial institutions and networked transactions.
Evidencing Performance
The company needed to prove that encrypted matching and analytics would detect criminal patterns at scale, not just in theory, so required an objective performance benchmark.
Syntheticr Ecosystem provided high-fidelity financial transaction and risk intelligence data, purpose-built for financial crime analytics, and accurately reflecting real-world conditions. This helped the team to:
Proof-Oriented System Validation
The Syntheticr scorecards served as an objective benchmark, so the company could prove where the solution worked and failed. This removed guesswork and accelerated product and engineering decisions.
Cross-Silo Resolution
Using multiple transaction datasets from different institutions, the team could easily test whether HE+ER was accurately processing and resolving fragmented data.
Encrypted Analytics at Scale
Syntheticr allowed realistic performance testing of HE+ER to confirm whether privacy-preserving analytics would work at real-world scale.
With Syntheticr Ecosystem, the defence company successfully developed and validated its HE-enabled entity resolution product for financial crime. The team gained clear evidence of detection performance, including the ability to uncover criminal behaviour spanning multiple data silos and that would not otherwise be identifiable.
Syntheticr Ecosystem empowers organizations to build and validate next-generation financial crime systems with realistic, labelled synthetic transaction and risk-intelligence datasets. Whether you’re developing encrypted analytics, multi-party detection, or cross-silo entity resolution, Syntheticr provides the trusted data foundation to train, test, and prove your solution at scale.
Test, train and benchmark your AML systems with Syntheticr