In the years since the 2008 financial crisis, non-compliance has cost banks over $321 billion in fines, settlements and enforcement actions.
In an effort to avoid fines, banks spend a staggering $270 billion a year on compliance—ten percent of their operating costs. And some estimates say that these costs could double by 2022.
With these budget-breaking amounts in mind, banks are strongly motivated to find cost-effective and easy-to-implement strategies to help them comply with regulations and avoid huge fines.
The Challenges of Compliance
Regulatory compliance is important for banks for financial reasons alone, never mind its supremely important role in stemming the tide of financial crime. Money laundering threatens national security by financing terrorists, human traffickers, drug dealers and other criminal activities. It stresses economies by diverting resources to combat such criminal activity, which slows financial growth.
There’s no question that financial institutions (FIs) understand the tremendous importance of regulatory compliance. So what exactly is it that makes compliance so challenging for banks? Why is complying with know your customer (KYC) checks—according to a report from Forrester Research—cited as among the biggest pain points faced by bank execs worldwide?
Some of the biggest compliance challenges faced by FIs come at the very beginning of the process, during onboarding. Here are two primary reasons why dependable compliance solutions can be tough for banks to implement:
- Onboarding processes are manual, time-consuming and labor-intensive.
- It’s hard for compliance programs to keep up with ongoing regulatory changes. Constant updates can be hard to adjust to in a timely manner, and flagged accounts negatively impact customer experience, causing extra work and stress for financial professionals and their customers. Which brings us right back to the difficulties of onboarding. Onboarding is a months-long process under the best of circumstances. Continual updates to regulations can drag this process out even longer.
Ahead we’ll share some of the best current solutions for the detection and monitoring of illicit criminal activity. There’s a bit of overlap with some of the capabilities in these different technologies, but that’s not a bad thing. Many of them enhance or compliment each other well.
Tech to the Rescue: The Best Fraud-Detecting Technologies
- Graph database technologies are systems that bring different information sources together into a single view of data, creating a comprehensive overview of each customer. A single data model makes it easier for compliance teams to identify patterns of illicit customer behavior. These technologies can also help FIs to find connections like those to politically exposed persons (PEPs) and ultimate beneficial owners (UBOs).
- Network analysis and visualization technologies depict connected entities as nodes—which are data points—and links that show their connections to other business entities through time. This allows financial institutions to see complex relationships between various people and entities laid out clearly before them—relationships that might easily be missed without the clear cues provided by these visualization technologies.
- Contextual computing uses hardware and software to analyze data about a device’s origins. As per the FINCEN advisory, the collection of data such as Internet Protocol (IP) addresses with timestamps, virtual-wallet information, device identifiers and other information for inclusion in SAR filings can be very helpful to law enforcement in identifying criminal entities.
- Predictive analytics uses statistical techniques like predictive modeling, machine learning, and data mining to analyze facts, past and present, to make predictions about future events.
- Big data is part of a data management strategy using advanced technologies to store, analyze and otherwise leverage massive amounts of customer data. These ecosystems of data are providing the ability to gain insights into potential illicit behaviors in ways that previously were not possible.
- Artificial intelligence (AI), particularly machine learning and natural language processing: Artificial intelligence is a branch of computer science that develops systems that perform tasks that would ordinarily require human intelligence.
Machine learning uses algorithms and statistical models that enable computers to perform tasks without being specifically programmed to do so. Together with natural language processing, which allows computers to process and analyze natural language data, these technologies give FIs all the advantages of having a bigger workforce without the effort and expense.
AI is the Technology that Supercharges All the Others
This brief overview reveals what these technologies all have in common. In their different ways, they each reveal connections and patterns in data that allow FIs to see businesses and individuals within a context.
Context quickly and clearly reveals conflicts of interest and deviations from normal behaviors that might otherwise be overlooked. This is where AI augments the capabilities of other technologies. False positives are a huge problem for compliance programs, with 98% of alerts never even resulting in suspicious activity reports (SARs).
AI uses machine learning to add a more human element that amounts to something very close to intuition. It uses that “intuition” to make decisions based on connections and patterns that it detects, greatly reducing the rate of false positives. PayPal has cut its fraud false-positive rate by half since it began using deep learning (a branch of machine learning that can understand images, sound and text).
For banks, adding AI to the compliance arsenal is like hiring thousands of human beings that all happen to have exceptional instincts. And machine learning enables AI systems to grow their intelligence with every bit of new data they ingest.
AI’s supercharged ability to identify trends without getting distracted by the minutiae is an unbeatable combination that could potentially save banks billions over the next ten years.
