Compliance as a Competitive Advantage <\/p>\n\n\n\n
A.J. Bosco<\/p>\n\n\n\n
The advent of data discovery software tools enables companies to discover information contained within it structure and unstructured data (\u201cBig Data\u201d) through Artificial Intelligence or Machine Learning that was formerly impenetrable. The evolving ability of AI to extract specific information from Big Data provides compliance departments (especially within the financial industry) with the ability to identify misconduct by employees and customers before it happens, thereby potentially saving companies from regulatory scrutiny, criminal prosecutions, large fines and the attendant reputational consequences. The challenge is collecting and analyzing structured and unstructured data across silos and adopting an enterprise-wide approach to data analysis. Companies that embrace the use of AI to Big Data for regulatory compliance will have a competitive advantage over their competitors both by forging closer working relationships between Compliance and the Front Office and by identifying potential problem employees and customers in or close to real time. <\/p>\n\n\n\n
Compliance traditionally works reactively, that is, examining events after they occur, often looking at one discrete activity at a time. For example, in monitoring for excessive mark-ups on fixed income trades, Compliance takes transaction data supplied by the front office, puts it through a filter, flags potentially problematic trades and hands the results back to the front office for action. Or the Front Office delegates functions, for example, email reviews, to \u201cdesignated supervisors\u201d, who conduct manual reviews or apply simple lexicons to identify emails that might indicate misconduct. By being reactive and looking at different transactions separately from each other, Compliance Departments do not efficiently access risk. <\/p>\n\n\n\n
But recently, risk and compliance functions in some companies have started to use enterprise-wide Big Data analytics supported by technology to help identify bad actors (both employees and customers) before they commit acts that cause financial and reputational risk. <\/p>\n\n\n\n
Examples of enterprise-wide structure and unstructured data being analyzed using AI to develop accurate profiles of employees can include:<\/p>\n\n\n\n
Trader misconduct can range from the minor rules violations (reporting trades late) to major misconduct (trading that manipulates markets and price fixing that violates antitrust laws). By using IA to aggregate and analyze some or all of the data points listed above for all traders on a particular desk, Compliance can draw a more complete risk profile of each trader and the desk as a whole. By using AI to analyze Big Data to produce these profiles, Compliance can proactively monitor for abnormal activity at or near real time. Higher risk profiles indicate a greater chance of misconduct, which in turn, would lead to putting those with higher risk profiles under enhanced scrutiny. <\/p>\n\n\n\n
Many financial institutions use static data bases consisting of \u2018know your customer\u201d information collected from customers at account opening. They compare this information against costumer transactions to try to identify if particular transactions are outside the customer\u2019s normal activity, which may need to be escalated for follow up. <\/p>\n\n\n\n
By applying machine learning to a single database comprised of information obtained at account opening and data gathered from the sources listed above companies can create more sophisticated and detailed customer profiles. This, in turn, can be used to generate behavioral analytics, which will improve KYC compliance by more quickly and accurately identifying potentially suspicious activity. <\/p>\n\n\n\n
The types of monitoring suggested in this article are not without potential problems. First, a number of jurisdictions outside of the United States will not permit companies to collect some of these types of data or permit it to be removed from the jurisdiction. Other jurisdictions have warned of the need for privacy regimes especially for Big Data. Additionally, holding out Compliance as a competitive advantage might be seen by prudential regulators as a blurring of the line between the front office and Compliance. The Federal Reserve and Office of the Controller of the Currency view Compliance as a line of defense against business misconduct and may look askance at a model that brings the two closer together. The question is whether regulators will applaud Compliance innovation or frown upon a control that doesn\u2019t conform to their expectations and the guidelines in their examination manual. <\/p>\n\n\n\n
After a period of rapid expansion, Compliance Departments at financial institutions are no longer growing and in some case are beginning to shrink. They will be asked to do more with less. They will also be asked to work as hard as ever to justify budgets and prove their worth. By exploiting data their companies already have on hand and applying IA to it, which admittedly, comes with a significant up front expense, Compliance Departments not only can become more efficient but will be able to add to the bottom line, or at least not subtract from it. If Compliance Departments can in fact prevent misconduct from occurring they will save money and put their companies at a competitive advantage over companies that do not adopt such methods.<\/p>\n\n\n
\u00a01. Structured data is <\/span>data is easily searchable by basic algorithms. Examples include spreadsheets and data from machine sensors.\u00a0 Examples of unstructured data include emails, text documents (Word docs, PDFs, etc.), social media posts, videos, audio files, and images. [site]<\/span><\/b><\/p>\n