The Real Cost of Government Fraud Is Measured in Lost Service

Author: Dave Stroth, Area VP of U.S. SLED at Elastic

Fraud is often discussed in terms of waste and improper payments, but that framing does not fully capture the consequences. Public programs operate with finite resources. When money is diverted through fraud, government agencies have less capacity to deliver benefits and critical services, like healthcare, nutrition assistance, unemployment benefits, or disaster relief.

Between $233 billion to $521 billion is lost to fraud each year across federal programs and operations. The impact is significant in programs designed to serve people during periods of hardship or crisis. Fraud can consume funding, time and investigative resources that would otherwise support legitimate beneficiaries. It can erode public confidence in programs that depend on trust to function effectively.

In addition, states are about to be financially penalized for fraud and error they structurally can’t see coming. The same data-fragmentation problem that led to billions in improper payments is about to convert into a direct hit on state budgets, and therefore on SNAP recipients, starting in FY2028.

Fraud is such a problem that President Trump’s administration established a task force to eliminate waste and the Department of Justice created a national detection center, working with agencies to combat fraud.

On top of that, bad actors are utilizing AI tools to scale fraud faster than traditional detection can keep up, using Generative AI to create synthetic personas at scale. To counter this, agencies are adopting modern AI capabilities to automatically triage alerts and surface subtle anomalies across datasets, empowering human investigators to focus on high-risk threats rather than sorting through noise.

Fraud exploits the gaps between government systems

Government agencies collect immense amounts of information that can expose fraudulent activity, but it sits across separate systems, programs and jurisdictions. A suspicious transaction may appear isolated while related activity exists elsewhere under a different identity, account or program.

Scammers exploit these gaps with multichannel schemes that outpace investigations. Agencies often have the evidence, but investigators can’t connect it fast enough to see the bigger picture or pattern.

For example, one bad actor committing unemployment insurance fraud, misrepresenting SNAP income and filing a fraudulent tax refund looks like three unrelated problems across three systems. Unified, it reads as one clear pattern.

Agencies gain incredible value by improving how authorized teams search, analyze and connect data. That requires appropriate governance, privacy safeguards and clear rules for how information can be used. This process gives investigators a complete view of suspicious activity while keeping within the legal and operational boundaries that govern public data.

Earlier detection preserves more resources for legitimate recipients

Traditional fraud enforcement occurs after money has already left a program. Agencies investigate activity, document what happened and attempt to recover funds. That remains essential, but full recovery is difficult.

A unified data approach creates a single operational picture by breaking down silos, and bringing together information from across systems, providing the context needed to identify suspicious activity before losses grow. When agencies can identify unusual relationships, conflicting information or patterns across transactions earlier, investigators have a better opportunity to intervene before fraudulent payments become significant.

Earlier detection can also reduce the burden on investigators by helping them prioritize the cases that present the greatest risk rather than sorting through volumes of disconnected alerts.

Agencies must turn existing information into insights that lead to action quickly enough for government leaders to make informed decisions.

Prevention must protect access. Many public programs serve people when timely help matters most, and false positives delay services and burden the very people programs exist to protect. Effective detection requires precision. A risk indicator should trigger review, not serve as proof of wrongdoing, and human judgment should remain part of every consequential decision. That includes monitoring false positives and understanding why cases are flagged.

Agencies should measure success on two fronts: fraud caught and friction avoided. Program integrity is strongest when agencies stop fraud without making services harder to reach for eligible people.

Getting started doesn’t have to be complicated

Rather than building massive new databases or executing costly ‘rip-and-replace’ overhauls, agencies can establish an AI-ready data foundation for fraud detection that unifies visibility across disparate IT environments while respecting agency data boundaries.

A practical first step is to connect and standardize the data and systems agencies already have. This includes:

• Establish data-sharing agreements between programs and departments
• Audit existing platforms to see if they can serve as a unified foundation
• Using entity resolution to connect records across systems, even when names, addresses or other identifiers don’t match exactly
• Setting up automated alerts for high-risk patterns using capabilities that may already exist within current platforms (often already supported, rarely fully configured)

Agencies can start small. Focusing on one program area and high-value data sources can produce results faster than trying to connect everything at once. Those early successes can then provide the experience and confidence needed to expand the approach over time.

Modernizing fraud detection relies on a better data foundation that enables data utilization, operational context, earlier intervention, and precise investigative workflows. State and local government leaders need a clear framework to guide this transition: search across legacy silos, observe high-risk anomalies in real time, and protect public dollars before they leave the building. Agencies that build that foundation stay ahead of evolving fraud, meet oversight requirements, and prove they’re good stewards of public funds.

The government must safeguard taxpayer resources, but the reason extends beyond the balance sheet. Every improper payment stopped is a dollar that stays available for the people these programs exist to serve. And for the states facing FY2028’s SNAP payment-error penalty, cutting errors now is the difference between funding benefits and funding a penalty that could run into the hundreds of millions of dollars.