AI-Powered Fraud Detection in Healthcare

A Fortune 50 health insurance provider wanted a better way to classify and detect potentially fraudulent claims, enabling their fraud investigators to focus their limited time on the highest-value cases for recovery. The company collaborated with CapTech to develop an AI-driven system that could quickly analyze vast amounts of data and flag anomalies, while assigning a risk score to prioritize claims with the most dollars at risk.

Fraud, waste, and abuse cases cause significant financial losses for health insurance companies, and the path to recovery can be cumbersome, requiring detailed analysis of vast amounts of claims data. The effective use of advanced technologies, such as artificial intelligence (AI) and machine learning (ML), can bolster claims review capabilities by swiftly identifying irregularities, reducing financial losses and enhancing operational efficiency.

When a Fortune 50 health insurance provider faced challenges in identifying and preventing fraud, waste, and abuse, it engaged CapTech to devise a tech-forward solution. The processes previously in place were query-based and required time-consuming manual review, leading to delayed responses and increased financial losses. The company needed an automated and intelligent system that could efficiently and accurately identify high-value claims for investigation.

The CapTech team began by working with the healthcare company to collect and integrate structured and unstructured data from various sources, including member information, provider information, and claims data.

By leveraging machine learning approaches focused on anomaly detection, the team developed predictive analytics capable of identifying subtle deviations from expected norms and flagging potentially fraudulent or wasteful activities. The team worked within an AWS cloud environment leveraging tools such as Amazon SageMaker, AWS Glue, and other cloud-native orchestration tooling to deploy the final solution.

The result of this collaborative effort is a robust analytics system capable of delivering real-time insights, empowering the healthcare company to proactively combat fraud, waste, and abuse, while optimizing operational efficiency and cost savings.

Once live, the system identified approximately $30 million in at-risk dollars, classified as potential overpayments that could be pursued through recovery efforts. Additionally, efficiency increased dramatically, with a 60% decrease in model delivery timelines.

As part of the project, the CapTech team also conducted a multi-day, cross-functional analytics training and ideation workshop. The workshop resulted in five new, actionable ideas for claim prioritization, and four new claim line processing concepts, as well as eight other ideas the company’s team was able to begin implementing.