PROJECT

Fraud Detection using Data & AI

CLIENT

Maryland Health Connection

DATE

AUGUST 2023
Case Study: AI-Powered Fraud Detection for PRAC (Pandemic Response Accountability Committee)

We partnered with the Pandemic Response Accountability Committee (PRAC) to implement an AI-powered fraud detection system, ensuring the transparent and accountable distribution of pandemic relief funds. Our solution leveraged advanced data analytics and machine learning to detect fraudulent activities in real-time, allowing PRAC to respond quickly and mitigate risks. By enhancing their oversight capabilities, we helped PRAC reduce financial abuse and ensure relief funds reached the people and businesses most in need.

The Pandemic Response Accountability Committee (PRAC) is a critical oversight body tasked with preventing fraud, waste, and abuse in the disbursement of more than $5 trillion in federal pandemic relief funds. PRAC monitors a range of relief programs, such as the Paycheck Protection Program (PPP), Economic Injury Disaster Loans (EIDL), and unemployment insurance benefits. PRAC’s mission is to ensure that these funds are used properly and benefit those affected by the COVID-19 pandemic, maintaining transparency and accountability across federal agencies and relief programs.
PRAC faced several challenges in safeguarding pandemic relief funds from fraud:
  • Enormous Data Volumes: With millions of transactions flowing through pandemic relief programs, manually detecting fraud was infeasible.
  • Complex Fraud Networks: Fraudsters used sophisticated schemes to exploit relief programs, making it difficult to detect patterns of abuse with traditional tools.
  • Cross-Agency Data Silos: PRAC had to oversee data from numerous federal agencies and states, which made identifying and linking fraudulent activities across systems challenging.
  • Overwhelming False Positives: Traditional fraud detection systems generated excessive false positives, slowing down investigations and misdirecting valuable resources.
We automated the data ingestion/transport mechanism, and developed a comprehensive AI-based fraud detection solution with in-house data scientist collaboration  tailored to PRAC’s unique needs. Our approach focused on integrating data, using AI to detect fraud patterns, and providing real-time actionable insights. Here’s how we tackled the challenges:
  1. Data Integration and Harmonization: We began by consolidating data from federal relief programs such as PPP, EIDL, and unemployment insurance, ensuring PRAC had a unified view of transactions across multiple agencies.
  2. Machine Learning for Fraud Detection: We employed machine learning models capable of analyzing vast amounts of transactional data to detect unusual patterns indicative of fraud. These models were continuously trained on evolving data to stay ahead of new fraud techniques.
  3. Minimizing False Positives: By fine-tuning the AI models and implementing advanced filtering mechanisms, we drastically reduced the number of false positives, allowing PRAC’s investigative teams to focus on genuine fraud cases without being overwhelmed by irrelevant alerts.
  4. Scalability and Adaptability: The system was designed to scale as PRAC’s oversight responsibilities grew, ensuring that it could handle the increasing volume of relief transactions and adapt to new fraud tactics.
  5. Collaborative Oversight: We facilitated seamless data sharing between PRAC and other federal agencies, allowing for more effective collaboration and faster identification of fraud schemes spanning multiple jurisdictions.
Group 2608703

Cloud migration and automation

Group 2608704

UX Experience Design

Group 2608705

Test Automation with Gen AI

Related Case study
Group 2608703

Cloud migration and automation

Group 2608704

UX Experience Design

Group 2608705

Test Automation with Gen AI

Menu

We provide
end-to-end IT
Services for Your Business