Webinar in Review
Self-Funding Project: Applying Robotic Process Automation in AML & FIU Operations
Learn how RPA can increase efficiency, lower risk, and trim your overhead
Date: October 26, 2018
By: Anshul Arora, Director, Head of FL Center, Matrix-IFS
Every financial institution faces costly challenges managing fraud and AML. False Positives (FP) are an especially problematic resource drain, as they typically range as high as 75 to 90percent. Since each alert must be investigated, the waste of time, technology, manpower – and budget – is staggering. What’s more, every FP potentially delays the detection and reporting of bona fide alerts and adds
“Over half the time our investigators spend on an alert is simple data gathering.”
– Top 10 US Retail Bank
Although the problem is industry wide, there are no standards for streamlining the process. Each institution is on its own to gather, sort data and investigate via external systems, many of which are decentralized.
With no decision standards, non-technical operations staff rely on manual tasks and calculations, creating inconsistencies, errors and auditing gaps. This disfunction between people, products and processes creates tedious work that leads to high turnover and higher training costs.
Applying Robotic Process Automation (RPA) delivers a fresh and efficient solution for financial institutions facing this costly problem. By utilizing software with artificial intelligence (AI) and machine learning capabilities, RPA enables processing of high-volume, repeatable tasks with far greater speed and accuracy than manual processing. Industries with complex data gathering and processing needs such as insurance, healthcare and manufacturing are already adopting efficient RPA solutions from a variety of vendors.
RPA makes sense where tasks are:
- Repetitive
- High volume
- Manually performed
- Event-triggered
- Require documentation of task-related decision points
A case in point is AML Financial Investigation Unit (FIU), which faces processing similar to traditional financial institutions:
- Predefined decision points
- Predefined investigation workflow
- Triggered when laundering scheme suspected and/or confirmed
- Heavy manual interaction with system
- AML alert volumes are higher than ever before
Evaluating and assessing FIU procedures uncovered multiple opportunities to streamline many steps through automation, including:
- Internal database systems
- External RFI systems
- Customer historical information
- Auto SAR e-filing
- KYC verification
- One-stop ecosystem
- Cognitive suggestion
RPA Can Be a Self-funding Project
In operations, RPA builds automated data collection and availability. It becomes tactical by applying smart data representation, scripting and macros to operation procedures. Providing BOT interfaces and information-sharing BOTs makes it scalable. Over time RPA becomes cognitive as BOTs get smarter, offer suggestions, and deliver quality analysis to fulfill your Compliance ‘Dream Vision’.
If well designed and correctly applied, RPA can provide significant and immediate operational improvements in accuracy, workflow, resources, and budgetary impact, including:
- Cut FP alert volumes
- Reduce investigation intensity
- Lower manual task (operations) volume
These results apply directly to the bottom line, meaning that Robotic Process Automation can be a self-funding project, maximizing both efficiency and ROI.
RPA Solution vs. RPAService – Which is Right for Your Institution?
A successful self-funding PRA project must be carefully initiated, undertaken only after comparing short- and long-term outcomes and ROI. At first glance, a vendor solution appears advantageous because it offers a potentially faster delivery and more standard training. Features may include:
- BOTs development framework
- Pre-built external data integration connectors
- Integration layer rather than UI
- Requires core product knowledge & support
- One-size-fits-all
Although service delivered by industry experts will likely take longer and up-front resources may at first appear greater, this approach provides immediate and long-term benefits. The time invested in customizing RPA to fit each institution’s unique internal infrastructure, organization, and training needs can deliver highly-efficient and productive results from the beginning. In addition, the resulting resource will readily adapt to future needs.Choosing to develop RPA specific to your infrastructure, workflow and needs provides uniquely flexible features, including:
- Builds BOTs using core technologies framework
- Customization to complement the product
- Bespoke configurations
- Covers integration and interface layers
- Smooth future case management upgrades
- Uses common languages (Java, XML, etc.)
Enhanced Operations Process Streamlines FP Identification and Investigation
Applying customized RPA speeds review of detection system processes by incorporating a multitude of data sources, all managed from a user-friendly single-screen dashboard.
Because a customized RPA can significantly reduce the volume of resource-wasting false positives, it enables investigators to identify and investigate high-probability alerts in a timely manner. Strategic alert dashboards can further enhance operational efficiencies by providing investigative support information.
Sample Alert Investigation Dashboard enables Investigator to view comprehensive information within the ecosystem and to initiate alert/case review immediately.
Self-Funding Project – U.S. Commercial Bank Before and After Applying RPA to Investigations
Recently a U.S. commercial bank factored operational efficiency before and after applying RPA to FP investigations. The results were significant and immediate.
Before applying RPA to processing False Positives, this bank had 50 dedicated employees averaging 400 alerts (at 30 minutes each) and investigating 250 cases (at 48 minutes) per day. RPA enabled this bank to boost productivity, dramatically cut FP volume and significantly increase investigations with the same 50 employees.
Applying customized Robotic Process Automation delivered dramatic results for the institution. Average alert investigation time was reduced from 30-minutes to 10-minutes, enabling the same team to process 1,200 alerts daily – 800 more than was previously possible. Case investigation time dropped from 48-minutes to 15-minutes, enabling the existing team to process 800 cases daily – 550 more than before. While investigator head count remained the same, faster processing and timeliness of investigations enabled more efficient tracking and identification of bona fide fraud and money laundering alerts.
Where to Start – Steps to Design and Implement a Self-Funding RPA Project
Although one-size-fits-all solutions can sound appealing, they lack full functionality and are hard to fit within existing infrastructure, and may not adequately address end-to-end needs or future-forward functionality.
The more carefully you evaluate, test and launch, the more beneficial RPA can be to your institution. These three key aspects must be investigated and addressed carefully up front.
Matrix-IFS takes a three-step approach to implementing RPA to ensure that each institution can maximize efficiency and effectiveness of existing infrastructure to best support current and future compliance and operations.
The best way to determine the value RPA can bring to your institution is to run a pilot project. Let the numbers speak for themselves. If you invest the time up front to carefully plan and deploy a bespoke pilot RPA project, you can drive both operational and economic benefits of significantly improved efficiency and accuracy for your institution.
Anshul Arora is Director of the Matrix-IFS Delivery Center, applying Subject Matter Expertise to help clients adopt and leverage new technologies for optimal efficiency. For more than a decade, he served as Program Manager, Architect and Specialist on more than 100 projects for major financial institutions and trading banks globally, addressing Compliance, RPA, Trade Surveillance, AML and Fraud models; and providing analysis, services and solutions on multiple platforms.