by Jerome Pugnet, GRC Product Marketing, SAP

Businessmen Looking at Computer MonitorsLooking at companies’ experiences, of various sizes and across all industries, I think we would all agree that fraud attacks often don’t come from where one would expect! Companies still rely too much on guess work and empiric methods while investigating potentially fraudulent transactions.

And to make things worse, fraud patterns evolve quickly and constantly. Thus, as companies put in place measures to prevent fraud, perpetrators quickly adapt and find ways to circumvent them. There’s clearly a need for better processes and tools to enhance their fraud detection and investigation.

Investigator Experience Isn’t Sufficient Anymore

To analyse and understand how and where fraud happens, one can’t just rely on the experience and intuitions of even the best investigators, or the analysis of standard fraud reports and basic metrics. Also, the more common analytical tools appear ineffective to scan very high and fast-growing volumes of data – where critical information to understand fraud patterns and hidden paths is buried.

Moreover, the range of data to examine to properly identify fraud trends is increasingly diverse – structured and unstructured. More than ever, fraud detection is a Big Data problem!

Fast Developing Predictive Technologies Offer Great Potential for Improvement

On the other hand, predictive analysis technologies are fast developing and becoming more widely available and powerful. They can help companies get deep insight into how and where fraudulent transactions originate, and analyse changing fraud patterns, in order to enhance their fraud detection strategies and adapt faster to new types of attacks.

This is why SAP, after recently launching our new fraud management solution, quickly integrated it with powerful solutions such as SAP Predictive Analysis and the brand new SAP InfiniteInsight solution from our KXEN acquisition.

This will help users improve the effectiveness of their investigations by better focusing efforts on higher fraud risks areas and continuously updating and refining their fraud detection strategies using predictive analysis data.

Predictive analytics solutions from SAP, that run on SAP HANA, provide needed capabilities as they analyse a wide variety and very high volume of data very fast. This helps to identify fraud patterns and trends: where fraud comes from, how it happens, what areas of the business it impacts, and so on. It’s also about going after the unknown and more complex patterns, and detecting earlier to minimize the damage of cleverly hidden suspicious transactions.

That’s not all – predictive analytics can also help fraud investigators by reducing false alerts resulting from inadequate fraud detection mechanisms: this is a critical issue today for many fraud investigators as they generate an excessive workload, and wasted efforts.

Can Predictive Analytics Benefit a Wider Audience?

The innovation brought by predictive analytics touches many other areas of the business, and in areas such as governance, risk and compliance (GRC), its use will develop to enable better predictability of risk, increased insight in areas of control weakness, support for internal audit programs, and so on.

These multiple applications create a high demand for experts such as data analysts and specialized business analysts, but the scarcity and high cost of these resources pushes for better usability of the tools. In the area of fraud in particular, invaluable expertise resides within fraud investigation teams who don’t have these skills as their primary asset.

For them, and others, it’s important that new predictive technologies become approachable for the non-experts, and more readily consumable by their most interested audience. This is one of the strong points of SAP InfiniteInsight, which helps reduce the reliance on data analysts.