data science (chombosan/


One-to-one unemployment insurance fraud investigations are over. Here’s what’s next.

Imagine for a moment that you’re a fraudster, and unemployment insurance (UI) fraud is your specialty. As the pandemic took hold in the U.S., you watched as unemployment numbers spiked. You took note as the government proposed and passed massive emergency legislation to provide unemployment benefits to millions of those put of out of work. You saw that the dollar amounts involved were enormous. The best part? Due to the contagious nature of the virus, everything had to be done remotely -- not only did the government not require in-person signings or verifications, it eliminated them altogether. 

This is a recipe for massive fraud -- a point that wasn’t lost on those with state- and federal-level responsibility for preventing and uprooting it. They knew this was coming; it was only a question of when and exactly how it would take shape. In short order, government leaders updated their unemployment fraud defenses as best as they could and got ready for the coming tsunami. 

That was in the spring and summer of 2020. Now with the benefit of some hindsight, we can see that despite their best defenses, state governments were simply unable to head off fraud at such a large scale. And as we settle into what may be a long haul of constantly fluctuating unemployment and continued uncertainty, it seems clear that a return to old ways of thinking about and managing UI fraud is not going to happen. 

What have we learned? 

So what have we learned in this period of expanded and accelerated fraud? Probably most important, it is now clear that the era of the “one claim, one investigation” approach to fraud management is over -- the scope and volume of fraud cases simply don’t allow it. Those charged with UI oversight need the ability to monitor large volumes of UI claims, registration and other types of data to identify patterns and signals pointing to potential fraud. Just as important, large-scale analysis can help uncover broader patterns that are simply unable to be identified on a one-to-one basis. 

What’s next: Beyond “one claim, one investigation”

Fortunately, this is not a new category of challenge, even if some of these analytics capabilities are relatively new to state governments. Banks, for example, have been deploying large-scale antifraud analytics strategies for years, with positive results. State governments can and should be putting “quick strike” capabilities in place that allow them to begin improving fraud detection in as little as four weeks, drawing from capabilities developed in adjacent industries facing similar challenges. Here are the specific capabilities state governments should be putting in place today, if they haven’t already:

Fraud vulnerability analysis

  • Use internal and external data to prep rapid UI data requirements.
  • Analyze existing processes to identify vulnerabilities.
  • Develop recommendations for vulnerabilities for existing and new fraud schemes.
  • Set action plans for remediating vulnerabilities.

Rapid detection platform

  • Analyze internal data across multiple unemployment programs, claims, employer accounts and more, using advanced analytical models.
  • Deploy link analysis capabilities for reviewing customer information and identifying clusters of suspicious activities.
  • Deliver daily reporting and analytics dashboards to help identify trends, patterns and signals.
  • Use scorecards to weigh analytical “scores” to prioritize investigative work.

Identity proofing and bot detection

  • Centralize development efforts -- including validation skills.
  • Deploy global operating model using lower-cost resources to achieve scale.
  • Standardize and streamline development, validation and governance procedures.

Integrated, real-time fraud detection

  • Embed event-driven fraud assessments into business processes to deliver real-time insights and results.
  • Direct workflow to the right groups or individuals in the organization using business rules.
  • Inform and protect the organization as new threats emerge, using cybersecurity scans.

If this list seems daunting, consider that states don’t need to have all of these capabilities in place at once to improve their ability to curb UI fraud. In fact, the above list goes approximately in order from least to most advanced, with fraud vulnerability analysis and a rapid detection platform being first-stage targets for many. Those capabilities can serve as the foundation for the other, more advanced capabilities.  

What’s most important is getting started -- expanding, improving and deepening existing UI fraud remediation capabilities. Because not only are fraudsters becoming more sophisticated, state governments are buckling down even further in the face of a revenue crunch. They are looking even more closely at massive programs such as UI to make sure they’re being run as efficiently as possible. And that begins with smarter approaches to fraud, waste and abuse.  

About the Authors

Luther Klein is a managing director at Accenture responsible for North American Finance and Risk Analytics.

Carl Hammersburg is manager of government and healthcare risk and fraud at SAS.


  • Records management: Look beyond the NARA mandates

    Pandemic tests electronic records management

    Between the rush enable more virtual collaboration, stalled digitization of archived records and managing records that reside in datasets, records management executives are sorting through new challenges.

  • boy learning at home (Travelpixs/

    Tucson’s community wireless bridges the digital divide

    The city built cell sites at government-owned facilities such as fire departments and libraries that were already connected to Tucson’s existing fiber backbone.

Stay Connected