The Simility Blog
How to Detect Fraud in an Omnichannel Banking Environment
Jayan TharayilNovember 15, 2017
As you are aware, banking customers today, especially mobile users, increasingly expect convenient access across channels, whether that means smartphone, tablet, browser, interacting with call centers, or even the old-fashioned approach of dropping by the branch. But did you know… 75% of consumers say that easily switching between channels when banking is important or extremely important, as is aggregated account access. Omnichannel is great for consumers. Here’s a surprising fact I found last week about how it affects your business.
Financial services organizations achieve a 9.5% increase in revenue year-on-year as compared to companies that do not take on an omnichannel approach.
9.5%! Wow. Needless to say, with happy customers and increase in revenue, an omnichannel approach is the way to go. But guess what? Fraud is omnichannel AND a booming business too. Fraudsters attempt to get past financial institutions’ various cross-channel authentication protocols. They are exploiting visible gaps in these authentication techniques and channels through strategies such as sim-swap, man-in-the-middle, phishing, keylogging, password guessing attacks or just buying the data on the dark net.
Then, they’re using the stolen credentials to execute their attacks. Recently, there’s been an increase in fraudsters targeting a victim’s bank contact center, and, through various means, coerces customer service to share basic account details. Then the fraudster accesses the account, adds a new payee and makes international funds transfer…and voila! The money is gone before it’s even detected as fraudulent. Then, this information is also used to access funds in other channels. In other cases, the attacks are carried out over a longer period of time, to allude traditional fraud velocity rules and technologies.
And all of your data is kept in either separate silos, or dumped into a deep, murky, data lake, where you cannot get real-time fraud insights across channels. But wait, it gets worse…It’s resulting in huge fraud losses for your company. (And I hate it when the bad guys get the goods!) And with the explosion of data and devices, these losses look pretty daunting. If legacy systems were working, this wouldn’t be happening.
What can adapt to devices, behavioral patterns, and evolving business, while reducing your fraud losses, preventing friction for good users, providing insights on user behavior while automating analyst workflow?
Let’s look at a case study that highlights what we did for a Fortune 1000 bank in North America. As a bit of a background, this company engages with its clients and prospects via multiple channels – and is engaged in offering multiple products (e.g. cards, retail banking solutions, loans, etc).
They are a forward thinking bank too. A few years ago, they realized that they had to break the silo-ed approach to fighting fraud. They were already working on an in-house fraud solution that was based on data lakes. However as they tried to grow, they ran into challenges around scalability, providing user-friendly workflows, and integrated machine learning.
Their vision was to create a single 360-degree view of the user across its various channels and products and also be able to apply advanced machine learning techniques which are specific to individual use cases they wanted to solve for such as account origination, account takeover, wire fraud and more.
In addition, the solution had to connect to various in-house investments – both legacy and new ones – like IBM MQ, Streamset.
SImility was successfully able to deliver on this promise and in the process also improved their response rates. The response rates varied from 25 msec on some channels to about 200 msec on other channels, while giving them a near real-time view of cross-channel events.