Financial Fraud Detection
Increase in successful fraud attacks YoY
Losses from credit card fraud in 2018
Cost to finserv companies for each dollar of fraud
Reimagine financial fraud prevention with Lucata’s software-agnostic solution. We enable deeper analytics on larger graph datasets than ever before without data pruning or database sharding. Lucata is compatible with open-source graph databases as well as custom-written graph engines built on the GraphBLAS framework.
The Financial Fraud Detection Challenge
Fraudsters are continuously getting more sophisticated. Fraudulent accounts look and behave much like legitimate accounts, making detection challenging. Detecting fraud requires going beyond individual account behavior, combining information from third-party sources, and analyzing relationships between many entities. Graph analytics using multiple data sources is the standard for detecting these behaviors. While this has made a difference, the analytics capabilities have not advanced far enough and financial fraud continues to result in tens of billions of dollars in annual losses. Current fraud detection solutions utilizing graph analytics are limited by the size of the graph database that can be analyzed, the depth of the analysis that can be performed, and the speed of the processing. Many of the current limitations can be overcome with better hardware to support higher performance graph analytics on larger datasets.
Graph database analytics for financial fraud detection are limited by the size of the dataset and the depth of analysis that is possible using existing hardware platforms. Lucata offers a solution to these limitations. Pathfinder supports graph analytics on unsharded scale 38 graph databases with fast analytics to more than 10 hops.
As an example, a top-tier consumer credit provider that loses billions of dollars each quarter from credit fraud can reduce their credit fraud losses by 90%. However, they need the ability to answer 6 questions from 6 different data sources (6 hops) in tens of milliseconds from at least a scale 30 graph store (230 data elements). Traditional graph analytics solutions are limited by their underlying hardware platform. Pathfinder provides a minimum 10x performance improvement and can support the analysis of previously impossibly large datasets, potentially saving hundreds of millions of dollars each quarter.
Lucata Enables Effective Fraud Detection at Massive Scale
The next generation performance you need to perform graph analytics on databases with 1 trillion nodes and beyond with no data pruning or database sharding. Detect potential fraud by digging deeper into your data than ever before possible.
Lucata Powers a New Era in Financial Fraud Detection
Financial fraud detection and prevention is more difficult than ever. Fraud solutions using graph analytics are in widespread use. However, current graph analytics solutions are limited in speed and scale by the underlying hardware. Lucata is compatible with open-source solutions and custom-written graph engines built on the GraphBLAS framework and can provide the speed and scale to solve the toughest fraud detection and prevention challenges.