Analytics & Big Data
Apr 26, 2019

Beyond the tip of the iceberg

Unraveling the complexities of financial-crime data management

In the fight against financial crime, false positives and investigation inefficiencies are some of the most persistent challenges facing financial institutions today. There's a reason these problems keep rearing their ugly heads. Underlying these challenges is the common problem of data management. So what impacts data, and how can financial institutions address the problem?

Three key data challenges

For starters, poor data quality can lead to bad alerts or false positives. Sometimes data is wrong. And sometimes data is missing, which can skew accuracy by its omission.

Second, there's the issue of disparate data sources. When data comes from different places it is often defined and measured differently in each source, which can lead to false positives. Today's investigators spend a lot of time collating data. Many pull up one screen for their case management system, only to minimize it and open a dozen different windows to collect all the data they need.

As a result, it can be difficult to ferret out whether an alert is legitimate. With no single 360-degree view of the customer to rely upon, due diligence becomes a time-consuming task. This is particularly challenging for larger institutions dealing in multiple jurisdictions or across many different product lines. And consolidation across the banking sector has only intensified the problem.

Finally, there's the issue of data volume. In some cases, particularly at larger banks, the sheer volume of data prohibits the transaction monitoring systems from scaling to it. And as part of the volume challenge, there's a lack of skilled resources to mine that data.

Improving data accuracy and availability

The solution, of course, isn't less data. In fact, data volumes will only continue to grow. Nor is the solution a third or fourth screen to access data that's siloed and spread across the organization. Instead, over the next five years, financial services will move more data to the cloud for data storage and manipulation. This may require new ways to think about governance or data management; it will also require new technical skills not typically seen in the relational database world.

But the journey doesn't stop there. The future will bring more adoption of artificial intelligence (AI), machine learning (ML), and natural language processing (NLP). This will be done first to automate data preparation and enrich data to improve the company's knowledge of its customers and ultimately beneficial owners. But tomorrow will also see AI, ML, and NLP used to extract business insight from the data by identifying complex patterns of potentially nefarious activity that people might miss using current rules.

Making investments more meaningful

Going forward, banks will invest heavily in building more coherent systems that can bring data pieces together. But they will still face the challenge of making these investments more meaningful.

Banks have access to all sorts of data being pumped out by bank customers on social media. And yet fraud continues to happen. The challenge is to pull all those pieces in and get access to all that data to make models more robust and drive better insights.

At the same time, we see a lot of Tier 1 institutions investing heavily in building data lakes and hydrating those lakes with data. We see many forward-looking institutions moving those data lakes to the cloud. But smaller firms will need to rely on turnkey solutions to manage data, track its lineage, maintain its integrity, and solve problems.

The need for a culture shift

Nearer term, banks have begun to pool their data by forming multiple consortiums. But this pooling doesn't come without cost. It's a value exchange – banks need to give data to get data. Thankfully, newer technologies allow banks to pool data in a way that favors a mutually beneficial exchange.

Perhaps most importantly, future success calls for a culture shift, one in which the enterprise and IT group are open to changing their philosophies around data management. While the traditional focus has been on data normalization first, the approach needs to change to one that is led by data consolidation – getting the data into a common place.

About the author

Satish Acharya

Satish Acharya

Global KYC and AML Practice Lead

Follow Satish Acharya on LinkedIn