Blog By Garima Chaudhary, Financial Crime and Compliance Management Specialist, Oracle Financial Services
As regulatory scrutiny increased, and monitoring systems started to generate an enormous number of cases, the option institutions had was to increase manpower to handle the workload. Increasing the manpower have drastically increased the cost of compliance. Such increasing costs have caused enterprises to start looking at advance analytics and data scientists as smarter investments in the long run. Analytics technologies like to predict upcoming cases. The technology helped reduce the false positive ratio by a large degree. These systems have remained mostly disparate in siloes, thereby decreasing the for most enterprises. Newer FinTech multi country-based products & services have further complicated financial crime risk making detection of organized, sophisticated crimes even tougher.
Traditional Approach – Ineffectiveness in Analytics
Added , need for machine learning, and , have been critical drivers for institutions to create sizeable highly skilled analytics team. Large data sets from production & non-production systems are needed for analytics to perform data discovery & regular model optimization. Generally, data provisioning can take very long, and by the time the data is made available, it might already be outdated. Moreover, drastic changes in the risk profile of customers or entities require access to non-production data along with production data. Again, a longer time for data provisioning does not provide must-have coverage to financial crime risk. Lastly, tools with restricted data science languages limit the ability to provide coverage for complex behaviors, such as a network of external & internal parties. This approach makes the data discovery & optimization process expensive & highly inefficient.
RPA Can Win Games but Not a Championship – Investigation Challenges
It is well established that, for cases, investigation analysts spend almost 80% of the time gathering the information and just 20% of the time analyzing the collected data. Time spent gathering data makes the investigation process the most expensive element in financial crime & compliance program. In recent years, has been leveraged to solve data gathering challenge. RPA has reduced overall data gathering time significantly, though it does not provide insight & connection between hidden parties. The lack of insight & connections makes the investigation process highly inefficient and exposes institutions to organized financial crime risks.
3 Ways to Turn Anti-Financial Crime Compliance into a Competitive Advantage
1. Unified Platform for Monitoring, Investigation & Analytics is Vital:
Quite a few leading institutions leverage a . However, analytics for data exploration, above the line tuning / below the line tuning and new risk coverage are in disparate systems. A unified strategic platform for analytics fully integrated with detection & investigation will ensure timely & quality availability of production data. Additionally, machine learning models require huge volume & a wide variety of data attributes. Data attributes such as case outcome & supported evidence can be efficiently fed into machine learning models, when programmed into unified fully integrated platforms. Additionally, a unified platform can drastically reduce the operational time of newly authored models to production, which is essential for the efficient model optimization process.
2. Polyglot Model Authoring with On-demand Data Access:
When a system is jammed with fewer analytics languages, it limits the institution’s ability to leverage numerous open-source languages as per specific requirements. For example, individual languages prove better while integrating with , or can better handle large volumes from Hadoop (Example: ) as compared to other words. Which implies, the ability to leverage multiple analytics languages is essential to the growing need for improved analytics. Strategic solutions should leverage popular data languages such as R, Python, and SQL, for productivity.
Institutions realize that active discovery requires the institution’s transactions, accounts, case, and other financial crimes related data hosted in a data lake. Ability to load data from the on-demand, significantly reduces the time and effort data scientists spend in preparing data for analysis. Data Scientists should be able to ‘mashup' production data with third-party data in the data lake for discovery and modeling.
3. Graph Analytics to Fight Organized Criminal Network:
Efficient Investigation of highly organized financial crime requires technologies such as to succinctly express intricate money movement patterns, detect multi-hop relationships, and identify hubs and spokes of activity. Graph Analytics leveraging a single source of data powers investigators with an ability to search customer information from various source systems and allows the linkage of customers, accounts, external entities, transactions and external data stored in disparate operational silos. A single source also provides a 360-degree view of a customer, foreign bodies, or account for a holistic view of the case, transactions, and external data of interest. Graph algorithms such as , can generate automatic linkages. For example, such linkages could be (i.e., linking based on customer identification numbers, name-matching, shared phone numbers, tax ID, etc.). Further, investigators can drill down (expand/collapse) on customer information & visualize related parties using graph analytics. Lastly, to auto-generates case narrative/summary for investigations, including case highlights, associated parties, number of events, and red flags, etc. can be a game-changer in documenting case findings.
Graph Analytics opens new avenues of deep learning using graph algorithms such as Graph Similarity. Graph Similarity involves determining the degree of similarity between Graphs. Intuitively, the nodes in both graphs would be similar if, its neighbors are identical (and its connectivity, in terms of edge, to its neighbors). Again, its neighbors are identical if their neighborhoods are similar, and so on. This intuition guides the possibility of using as a method for measuring Graph Similarity, precisely because of the nature of the algorithm and its dependence on neighborhood structure.
With seamless access to production data in a secure and designated discovery sandbox, ability to leverage popular data science languages and graph analytics, both data scientists & investigators can gain an accelerated path to explore financial crimes data & hidden unknown networks interactively.
In the upcoming , Oracle is hosting a panel of key industry experts. The Panel, , features - Global Head- Financial Crimes Risk Management, and Group Chief Anti-Money Laundering (AML) Officer, Scotiabank is one of the panelists. Stuart Davis continues to speak about his vision of a unified compliance data store for all financial crime needs. This data store is for monitoring, investigation & advanced analytics. Another panelist is - Vice President, Compliance Surveillance Technology & Analytics Group, Charles Schwab. Brad Ahrens is a massive advocate of using innovative technologies to improve advance analytics & intelligence productivity. from our partner organization PwC will also be part of the panel. Chief Financial Crime Consultant will moderate the panel, Oracle and (Jason)-[Loves]->(Graph). Come join us on September 24th between 3:45 PM to 5 PM to listen to insights on how using innovation, compliance investments can be turned into competitive advantages.
Oracle experts will also be addressing critical issues faced in fighting financial crime at Booth #200. We look forward to meeting you at the ACAMS 18th Annual AML & Financial Crime Conference between September 23rd and 25th in Vegas.
For more information on Oracle’s Financial Crime solution please visit our websites at:
Oracle Financial Crime and AML Compliance Management:
Oracle Financial Services: