Technology Blogs by SAP
Learn how to extend and personalize SAP applications. Follow the SAP technology blog for insights into SAP BTP, ABAP, SAP Analytics Cloud, SAP HANA, and more.
cancel
Showing results for 
Search instead for 
Did you mean: 
Former Member
In today’s digital age, there can be a contradiction between the need for more speed and real-time transacting on one hand, and the necessity to mitigate the risks of conducting business on-line. These risks include potential fraud, exposure to unsafe business partners (be they customers, vendors, service providers), failures to comply to industry or financial regulations, and so on.

How Is It Done Today?

To mitigate those risks, companies rely on technologies that can help them scan transactions for fraud or irregularities, perform compliance checks, or screen business partners vs public lists or other databases from private providers such as Dow Jones or Thomson Reuters, and raise alerts for them to investigate. The amount of data involved is growing exponentially so more and more power is needed—these are clearly Big Data types of requirements.

However, many still use traditional tools, often based on basic data mining or more elaborate business intelligence technology, which show their limitations and can involve a significant amount of manual work and turn out to be quite costly and unreliable. Also, many of these tools that used for anomaly detection and screening are not integrated with the business systems where the data that needs to be scanned reside, so the cost of customizations and maintenance can be significant.

More recent analytical offerings have increased the level of sophistication, leveraging web services or other more-or-less effective Extract-Transform-Load (ETL) technologies to tap into transactions, third-party data, and other records and perform fraud detection and screening. Pretty complex algorithms can be developed to allow them to detect more hidden patterns, or suspicious relationships, for example, but often requiring highly skilled (and scarce) data scientists.

But even the more sophisticated tools scan transactions or


new business partner  information after the fact.


In view of the very high volume of data processed during the busiest hours, and to avoid disruptions, typically mass screening is performed in batch at night, when there is less traffic. But this is not ideal as the pace of business is ever increasing, and because so much of it is now online and happening 24/7.

And as you are continuously on-boarding new customers and transacting with them online, it isn’t great if they’re found to be problematic in the next screening run and you’ve already engaged with them.

Live Business Needs Real-Time Protection

Obviously the speed of business and volumes of data isn’t going to decrease, and the use of the term “live business” is becoming more and more widespread. This also increases the level of risks and the frequency of occurrence of potentially fraudulent transactions, anomalies, compliance issues, or exposures to risky third parties.

This means that technologies that help detect these risks need to be in sync with the pace of the business—they need to operate in real time.

If we take the example of online merchant sites—the Amazon, Uber, Sky Media of this world—which add thousands of customers every day to their records, they need to be able to perform checks in real time to verify they are not on a risk-list or sending illegitimate identity or payment information, and this needs to occur before they start to transact with them.

The vast majority of customers are legit, so it’s important that these checks don’t keep them waiting so that they are not tempted to just give up or go to a competitor.

This requires technologies that are able to process these masses of information in real time, based on the latest innovations like:

  • In-memory computing that enables Big Data capabilities to process high transaction volumes extremely fast

  • Powerful data analytics and predictive technology that can analyse all different types of patterns and variety of data

  • Easy-to-use alerting and investigation features, so that alerts can be analysed and processed quickly and effectively by the right teams




All these are areas where at SAP we have invested intensively these past years, to develop solutions that can respond to these needs with the desirable speed and efficiency, and providing a very collaborative and highly visual user experience to help investigators.

But It’s Not Just About Speed—It’s about Accuracy

In this context of fast-paced business, accompanied by real-time risk detection tools, high accuracy is also a critical element. Without it, continuously screening high volumes of transactions and third-party data could generate very high numbers of alerts, many of which may turn out to be false positives, with the effect of burdening investigators and significantly reducing their effectiveness.

They need to provide simulation capabilities, for instance, to fine-tune detection rules and threshold, and include a degree of intelligence, for example, to recognize situations that had already been previously identified as false alerts, or to help improve detection on a continuous basis. Machine learning is the next big step, and it’s developing fast.

And It’s about Ease-of-Use

Fraud and other risk situations can be quite complex to identify, and patterns are constantly evolving, and although technology can increasingly help as I just mentioned, it can never totally replace the experience of practitioners and seasoned investigators.

As they seldom have the skills to manipulate complex algorithms, fraud and third-party risk screening tools need to be more approachable. This is accomplished by offering more intuitive, visual features to help them analyse, improve their rules and detection criteria, and add new ones to respond to ever changing patterns or new types of risks or compliance requirements (for example) as the business expands and evolves.

Benefits of Today’s Innovative Fraud Management, Screening , and Predictive Analytics Solutions

In conclusion, today’s best fraud management, screening, and predictive analytics solutions have many benefits. They can:

  • Process very high volume of data in real time and adapt to the pace of increasingly digital business, leveraging in-memory computing technology, advanced analytics, and intuitive, collaborative user tools



  • Identify fraud situations, irregularities or third-party risks not just quickly, but also more accurately, to minimize false positives and stay in control even in fast-paced commercial activities



  • Enable going after the less known and more complex patterns and networks, leveraging not only innovative technologies like predictive analytics, but also the experience of best-qualified practitioners




The simple result of these innovations is that these risks or issues can be detected earlier and more accurately, the potential damage and impact on revenue of cleverly hidden suspicious transactions or unreliable business partners can be dramatically reduced, without slowing down the business.

Find Out More…..