A government agency department responsible for investigating potential cases of customer fraud had a large collection of records about their subjects, but they wanted a better way to investigate new cases. The current process was inefficient and time-consuming, as they had to manually sift through paper records or manually search for records in multiple systems. They wanted a solution that could systematically ingest data from existing data sources and then visualize, explore, annotate, and record relationships between entities, including customers, their addresses or businesses, and historical activity. They wanted investigators to be able to develop cases digitally and collaborate with colleagues on them. The agency had a limited budget for technology and wanted a cost-effective solution that could be implemented quickly.
Federal Government
Data Management & Analytics, Custom Software Development, Risk and Fraud
React, Java, Spring Boot, D3.js, Neo4J, AWS
We worked with the customer to review their existing solution and manual processes to explore cases and recognized early on the need to provide two core capabilities: federated search and entity analytics and visualization. We leveraged an existing internal federated search product and extended it with capabilities to incorporate returned data into charts. The charts had to be interactive, allow panning and zooming, allow annotations on the chart or individual entities, enable collaboration between individuals or teams, track changes during chart development, and, most importantly, enable users to define automated analysis rules. These rules allowed users to visually identify potential risk factors such as records with inconsistent attributes or associations with unreputable businesses or associates.
By leveraging existing capabilities and a mature delivery pipeline, our team was able to deliver a working prototype deployed into AWS within three months. The working prototype integrated into multiple production systems and allowed users to visually explore data and relationships and validate the results against existing cases.
Working prototype in production environment in less than three months.
Streamlined process for investigating potential cases of customer fraud, eliminating the need for manual searching through paper records or multiple systems.
Created a systematic ingestion of data from existing sources that allows investigators to visualize, explore, annotate, and record relationships between entities.
We’ll work with you to deliver a product built on technological experience, the research know-how, with built-in measurable results.