For millions of Americans, public benefits are lifelines during some of life’s most difficult moments. Those lifelines are increasingly under attack. Fraud and improper payments siphon billions of dollars from essential services, while AI gives bad cyber actors faster, cheaper and more sophisticated ways to exploit government systems and the people who depend on them.
Public sector agencies now face an urgent challenge to stop fraud before the money is gone while ensuring legitimate citizens can still access the support they need. The scale of the problem is underscored by the fact that the federal government issued about $186 billion in improper payments in fiscal year 2025.
AI is fueling the fraud landscape across the globe. With GenAI as accessible as it is today, it’s easier to automate cyber attacks, impersonate others via synthetic identity theft or create realistic and large-scale phishing scams. Preventing fraud and protecting citizens are the same job, so the public sector needs proactive fraud strategies that move at the speed of AI.
Traditionally, the federal government’s fraud response has been retrospective. An agency detects improper payments in an audit and then tries to claw the money back after it’s gone. In 2023, GAO surveyed 24 federal agencies and found a third didn’t have regular monitoring or evaluation activities, and half didn’t regularly make changes based on evaluation results.
Agencies need capabilities in place to flag anomalies before any disbursement. This is a true shift from reactive tactics to proactive action and it will largely come through holistic data visibility.
Agencies Need to See Everything, As it Happens
Public sector agencies sit on unprecedented amounts of real-time and historical information but lack a way to find and analyze it holistically, especially when information lives in different systems, regions and formats. This information includes both structured and unstructured data, such as constituent and beneficiary information, claims data, application and transaction details, case files, home addresses and more.
This variety of disconnected data creates a challenge for program integrity and fraud teams who need holistic data visibility to detect patterns instead of mere one-offs within an isolated system or dataset.
To combat this, agencies should implement an integrated, platform-based approach that ingests and normalizes data from siloed systems into a single, searchable environment where anomalous claim patterns surface in real-time instead of months later.
For example, teams working to combat unemployment insurance fraud can apply custom detection rules across applications and payment data to identify fraudulent unemployment insurance, SNAP, or disability claims and flag or block improper payments before funds go out the door.
Securing Citizen Data is Fraud Prevention
Full visibility into data leads to stronger cybersecurity efforts. Stolen personal information is the raw material of benefit fraud. This data becomes the input for synthetic identities and claims filed in real people’s names. So securing citizen data and preventing fraud are two sides of one coin. A breach prevented is a fraud prevented.
California’s Employment Development Department, which administers benefits that serve millions of California residents, provides an instructive case study.
Through an AI-powered SIEM (Security Information and Event Management), CA EDD consolidates system and transactional data from nearly 3,000 servers, 14,000 endpoints and 10,000 employees into a single, searchable location, giving the team visibility to spot patterns and vulnerabilities across the entire environment.
The results have been positive, including a 99% reduction in mean time to response as well as 850 billion records secured and integration across 3,000 servers. This reduction shows that speed is the biggest difference between stopping a fraudulent payment and chasing it.
For agencies to truly face the threat of AI-enabled attackers and provide security to citizens who need it most, technology teams must unify data across silos and build explainable, auditable analytics that satisfy oversight and due-process requirements.
By pairing AI-driven triage with human expertise, and treating data retention and compliance as design requirements, the government can ensure citizen access to critical benefits. Every fraudulent dollar stopped is a dollar returned to the public and preserved for someone who genuinely needs it. Data-driven prevention helps agencies honor that obligation.
Jennifer Nowell is an accomplished technology sales executive with more than 25 years of experience driving growth across the public sector. Over the course of her career, she has built high-performing organizations from the ground up, pioneered modern sales strategies, and cultivated new teams across a range of public sector markets – establishing a track record of turning emerging opportunities into sustained business.
Throughout her career, Jen has led direct sales organizations serving federal, state, and local government, healthcare, and education customers. She now serves as Area Vice President for Federal Civilian at Elastic, where she leads the company’s strategy to deliver cybersecurity, search and data analysis solutions to federal civilian agencies.
Leave a Reply
You must be logged in to post a comment.