Follow Datanami:
August 16, 2022

Using AI to Automate, Orchestrate, and Accelerate Fraud Prevention

John Larson, Ernest Sohn, and Teresa Camacho


As fraud, waste, and abuse (FWA) cost government agencies and private companies billions each year, artificial intelligence (AI) has transformed how organizations prevent, monitor, and respond to FWA activity, powering advanced analytics, repeatable processes, and workflow automation and orchestration tools.

Indeed, AI offers a wealth of tactics for combatting fraud, such as advanced authorization for credit card transactions, deep learning to combat false positives, and behavioral analytics. However, here we explore another facet of AI’s fraud-fighting potential: helping all of the pieces work in concert towards a “fused” FWA solution.

An Expanding Fraud Environment Demands a Fresh Approach  

Fraud targeting and tactics have grown more sophisticated—and expansive—in recent years. Take fintech, for example. The rapid proliferation of apps, payment platforms, and digital assets entering the financial services ecosystem has often outpaced regulatory oversight, and many of these channels have become a magnet for illicit activity. In 2021, criminals laundered $8.6 billion through cryptocurrency, a year over year increase of 30%.

Meanwhile, fraudsters have been adapting their schemes and scams to developments in the geopolitical landscape. In 2020 and 2021, COVID-19 presented a national crisis, necessitating rapid government intervention through existing and new safety net programs. The size and speed of this governmental response created opportunities for criminal exploitation. In the federal unemployment program, for example, fraudulent activities quickly overwhelmed many state systems with an estimated $200 billion stolen from pandemic unemployment funds. In 2022, scammers were quick to exploit the war in Ukraine, posing as Ukrainian refugees, launching fake charity websites, and using several other predatory tactics to trick people into giving money via online donations, money transfers, and even cryptocurrency.

Fraudsters made off with more than $200 billion in COVID-19 funds (pd studio/Shutterstock)

Traditional approaches to fraud prevention and response no longer measure up. First of all, they’re reactive, rather than proactive, focused on damage that’s already taken place rather than anticipating, and potentially preventing, the threats of the future. The limitations of this approach play out in commercial off-the-shelf tools that organizations can’t easily modify to new developments in the landscape. Even the most cutting-edge AI solutions may be limited in detecting new types of fraud schemes, having only been trained on known categories.

Secondly, today’s siloed operations impede progress. Cybersecurity teams and fraud teams, the two groups on the frontlines of the fight, too often work with different tools, workflows, and intelligence sources. These silos extend across the various stages of the fraud-fighting lifecycle: threat hunting, monitoring, analysis, investigation, response, and more. Individual tools address only discrete parts of the process, rather than the full continuum, leaving much to fall within the gaps. When one team notices something suspicious, the full organization might not know about the threat and act upon it until it’s too late.

Bringing the Pieces Together with an AI-Powered Fraud Solution

The next-generation hub of security operations brings cybersecurity and FWA specialists together into a continuous loop of information exchange. A “fused” team hunts for threats across domains and shares data and surveillance technologies and case management tools, for more effective, proactive response than with traditional fraud-fighting approaches.

A successful fraud solution is comprised of:

  • A centralized operating model, in which internal and external stakeholders regularly coordinate, communicate, and share data;
  • Intelligence tradecraft, such as skilled threat hunters analyzing dark web forums and marketplaces, proprietary data sources, social media, and more to identify threats and anticipate attacks;
  • Integrated case management and continuous improvement, to ensure that fraud detection models are continually updated and responsive to changing fraudster behavior;
  • Advanced analytics and AI technologies.

The AI-Powered Fraud Solution in Action

One example is the integration of case management and analytics to automate workflows, repeatable processes, and feedback loops. Applied in this way, AI makes case management—traditionally conducted through manual workflows—faster and efficient, with an electronic “paper trail” of accountability throughout.

Advanced analytics and AI are key components of a successful fraud-fighting effort (U-STUDIOGRAPHY DD59/Shutterstock)

When analytic models detect suspicious transactions or activities, the AI-powered system can automatically open a case and route it to the right person for analysis, with chain of custody and data lineage maintained throughout. With automated and standardized workflows and business processes, examiners can triage cases based on available workloads.

Information on case outcomes automatically feeds back into the system, to help ensure that fraud detection models continuously update and evolve in response to changing fraudster behavior.

When fraudsters are found, analysis of their behavior patterns helps to fine-tune threat detection models even more, sharpening an organization’s ability to home in on future threats.

AI helps facilitate and accelerate detection across the fraud landscape as well. Improved analytics make it possible to attribute specific actions to specific actors, so when a fraudster’s identity becomes known, FWA fighters can map their digital footprint to networks and websites, potentially exposing other malicious activity.

Moving Your Fraud-Fighting Efforts Forward

To get started with implementing an AI-powered fraud solution within your organization:

  • Enlist a senior leader to champion the program;
  • Evaluate your current capabilities and benchmark them against peers;
  • Craft a strategy grounded in the common challenges, tools and indicators between the fraud and cyber teams;
  • Communicate your strategy early and often;
  • Gain traction and support across your organization through quick wins;
  • Keep data separate until you start understanding the types of fraud you’re seeing and processes to address them. Then start “fusing” this data a little at a time;
  • Establish a baseline to determine what success looks like: the type of fraud you are detecting, response times, number of fraudulent attempts blocked, etc;
  • Layer technology solutions to achieve your metrics, leveraging feeds and solutions already in existence;
  • Keep fraud detection processes repeatable and consistent;
  • Continuously evaluate and improve your efforts.

By establishing the processes and tools for a “fused” cyber and fraud team, an AI-powered fraud solution enhances an organization’s overall security posture, creates a more hostile environment for fraudsters, and strengthens the organization’s ability to anticipate threats. AI is an essential component to such a center, bringing tools, teams, and tactics together through automated processes, orchestrated workflows, and algorithms that are always learning from the most current, comprehensive sets of data.

John Larson

Ernest Sohn

Teresa Camacho

About the authors: John Larson is a senior vice president in Booz Allen’s Chief Technology Office, leading analytic strategy advisory services; fraud, waste, and abuse detection and mitigation; and artificial intelligence and deep learning services. Ernest Sohn is a director in Booz Allen’s data solutions and machine intelligence group. Teresa Camacho is a chief scientist and leader in Booz Allen’s digital, analytics, and strategy practice serving civil and commercial clients.