What it is and why it matters
Fraud prevention technology has made enormous strides from advances in computing speeds (high-performance analytics), machine learning and other forms of artificial intelligence (AI). Fraud touches every area of our lives; it raises the price we pay for goods and services, squanders tax money, pulls resources from innovation and even costs human lives.
Fraud can encompass waste and abuse, improper payments, money laundering, terrorist financing, public security and cybersecurity. In the past, organizations had to take a fragmented approach to fraud prevention, using business rules and rudimentary analytics to look for anomalies to create alerts from separate data sets.
Data couldn’t be cross-referenced through automation, and investigators couldn’t manually monitor transactions and crimes in real time; they had to do so after the fact. In health care, fraud prevention was more like “pay and chase”, because the criminal was long gone by the time fraud was detected.
To combat fraud, newer technology has been developed to predict conventional tactics, uncover new schemes and decipher increasingly sophisticated organized fraud rings. This involves more than standard analytics; it applies predictive and adaptive analytics techniques – including a form of AI known as machine learning. By combining big data sources with real-time monitoring and risk profile analysis to score on fraud risk, fraud prevention has evolved to start turning the tides of losses.
Fighting Identity Fraud with Analytics
Identity fraud is a growing concern that affects both businesses and customers. Fraudsters now have easier access to more tools and data than ever before, causing identity theft to reach a record high. This chart compares account take over, card not present and other forms of identity fraud losses and their growth, which continues to rise.
Fraud detection in today's world
The growing complexities of state-sponsored terrorism, professional criminals and basement bad guys are becoming harder to understand, follow, expose and prevent. Fraud detection in today’s world involves a comprehensive approach to match data points with activities to find what is abnormal. Fraudsters have developed sophisticated tactics, so it’s essential to stay on top of these changing approaches of gaming the system.
Many times, cybersecurity breaches enable fraudulent activities. Take for example, retail or financial services: Once a luxury, real-time transaction monitoring is now a baseline requirement, not only for financial transactions, but for digital event data surrounding authentication, session, location and device.
To identify and stop an array of fraud attacks and crime quickly and accurately – while improving customer and citizen experiences – organizations should follow four critical steps:
- Capture and unify all available data types from across departments or channels and incorporate them into the analytical process.
- Continually monitor transactions, social networks, high-risk anomalies, etc., and apply behavioral analytics to enable real-time decision making.
- Instill an enterprisewide analytics culture through data visualization at all levels, including investigative workflow optimization.
- Employ layered security techniques.
The fraud detection and prevention technology that you choose should be able to learn from complex data patterns. It should use sophisticated decision models to better manage false positives and detect network relationships to see a holistic view of the activity of fraudsters and criminals. Combining machine learning methods – such as deep learning neural networks, extreme gradient boosting and vector machines – as well as proven methods such as logistic regression, self-organizing maps, random forests and ensembles – has proven to be far more accurate and effective than approaches based on rules.
Cutting False Positives in Half
We’ve increased our fraud detection rate by 50 percent and reduced card fraud by 50 to 70 percent for cards under the optional prevention program — all while cutting false positives in half, this means that we’re not only detecting and preventing more fraud, but we’re inconveniencing fewer cardholders in the process.”
Who's using fraud prevention?
Business and governments alike have embraced technologies like data visualization and artificial intelligence to greatly reduce and even prevent the economical and reputational repercussions of fraud. Analysts and investigators work together, breaking down siloes, scoring and prioritizing alerts based on severity, then route high-priority alerts for more in-depth analysis.
Fraud is often perpetrated through synthetic identities, customer account takeover, nefarious applications, digital payments and authentication, procurement and other financial crimes. Financial institutions detect fraudulent transactions in real time with fewer false positives and detect money laundering or terrorist financing through complex algorithms looking at a multitude of factors.
Claims fraud runs rampant, and application fraud is on the rise. Instead of the pay-and-chase approach – after money has been spent – data analysts are preventing fraud by applying algorithms to detect anomalies and patterns. Analyzing multiple factors to determine how claims fraud is perpetrated, not only can fraud be detected when it happens, but more importantly, fraud can be prevented before it’s too late.
Governments are now combining siloed data to catch tax fraud, predict intrusions, identify abnormal behavior, and to shut down real-time and future threats. All of this work enhances border security, gathers intelligence for law enforcement, monitors opioid abuse and keeps children safe.
Health care claims fraud costs millions, even billions, worldwide. Health care organizations are successfully preventing fraud by taking an enterprise approach to payment integrity and health care cost containment by using advanced analytics.
Learn About Other Industries
How fraud prevention works
Fraud detection and prevention is not a
static process. There’s no starting and ending point. Rather, it’s an
ongoing cycle involving monitoring, detection, decisions, case
management and learning to feed improvements in detection back into the
system. Organizations should strive to continually learn from incidents
of fraud and incorporate the results into future monitoring and
detection processes. This requires an enterprisewide analytics life
Your goals may involve fraud detection, compliance or security. As technologies like artificial intelligence and machine learning have become more prevalent, the next generation of technologies is automating manual processes associated with combining large data sets and employing behavioral analytics
Supervised machine learning algorithms learn from historical data, identifying patterns of interest that an investigator might want to flag.
Unsupervised machine learning assesses and examines data that does not contain identified fraud. It is used to uncover new anomalies and patterns of interest.
Network analysis to identify paths, connections and hubs that reveal patterns and social networks of interest that are essential to an investigator’s toolkit.
“The rise of the digital economy has been matched by the rapid spread of fraud and cybersecurity risks. We want to meet customers where they are in their analytics journeys, particularly as they adopt technologies like AI, IoT and cloud. With SAS to help them, they will be better equipped to break down data silos, adjust to shifting regulations and safeguard against present and future risks.” Stu Bradley Vice President, Fraud and Security Intelligence Practice at SAS
Explore SAS Solutions for Fraud, AML and Security Intelligence
Featured product for Fraud Prevention
SAS® Visual Investigator
SAS Visual Investigator is a fraud detection, investigation and incident management solution that combines large, disparate, structured and unstructured data sources. Through a visual user interface, investigators can define, create, triage and manage alerts and perform detailed investigations to uncover hidden behaviors and activities.
- Article Next generation anti-money laundering: robotics, semantic analysis and AIAnti-money laundering taken to it's next level is sometimes referred to as AML 2.0 or AML 3.0. What does this next wave of AML technology look like? What can it do that you can’t do with traditional AML? See the results innovative financial institutions around the globe are already getting.
- Article Small-time cheats and organized crime: Benefits fraud re-examinedTo combat benefits fraud, better utilize resources and improve ROI, analytics can detect organized crime rings rather than just small-time fraudsters.
- Article Transaction monitoring: To segment or not?A typical anti-money laundering transaction monitoring program has scenarios that monitor the customers and accounts that pose the most risk. The fact is … this one-size-fits-all methodology isn’t very effective because customers transact differently based on many factors. So how do you incorporate that into your program?