By Stuart Brock
International efforts regarding anti-money laundering (AML) and counter-terrorist financing have intensified in the last decade; as governments and the private sector work to staunch the cross-border flow of funds used by criminal and terrorist organizations. As a result, bad actors have become more sophisticated in their methods, moving their activities away from the global financial system and into the murky world of trade-based financing.
While other industries have adopted data-driven frameworks and technologies, which are well suited to monitoring and revealing wrongdoing, trade finance remains remarkably document-based. Paper contracts and agreements are generally reviewed manually to ensure compliance with regulators. In cases of concern, contracts are individually examined by specialists – who, like first-line reviewers, are prone to human inconsistency, error, and delay.
Banking and financial services firms are now deploying artificial intelligence (AI) to analyze contract data, as a means to identify cash smuggling and financial system misconduct under AML laws and regulations. These institutions are using this technology to determine if their supply- and sell-side contracts fully align with internal risk processes and regulatory obligations. While applying AI-based reviews for ongoing end-to-end assessment of third-party paper. This entails more than automating the examination of documents or data, but determining if these instruments define the systems, processes, relationships, and the controls necessary to maintain AML transparency.
Legacy processes undermine compliance
Criminals often employ invoicing tactics, such as over- and under-pricing trade goods, so that an importer can resell goods and recover the difference in fair market value on the open market; issuing multiple invoices to achieve duplicate payments and misrepresenting goods in terms of value or quantity, so that importers can recover the difference in cost.
Criminal entities also conduct transactions across poorly regulated markets, particularly in developing countries, where infrastructure is weak or non-existent, or across free-trade jurisdictions, using shell companies and bank accounts that add legal complexity to trade finance.
In these institutions, the quality of the data represented by contracts erodes the ability to standardize manual document review, and fragmented data repositories inside and outside corporate boundaries make aggregation difficult to achieve. Examiners spend a disproportionate amount of time and effort finding contracts and extracting the data and legal clauses they contain. This critical, but methodical task, greatly reduces the time available for their investigative efforts as they focus more on the process than the result itself.
Moreover, manual methods for determining contractual risk and scoring of that risk are notorious for producing false positives. Out of an abundance of caution, investigation of low-risk contractual documents is a drain on resources and diverts attention from the much more significant sources of money laundering and trade finance violations. As mandates become more complex and numerous, additional evaluation criteria and divergent thresholds only serve to compound this problem.Inconsistent standards for contract language require a great deal of human interpretation, with screeners often disagreeing on what constitutes risk and violation of AML compliance. Qualitative and quantitative metrics for assessing risk across jurisdictions and processes are often equally inconsistent.
Certainly, strides in operational efficiency and contract governance can improve trade-finance crime-prevention. Firms are making strategic investments into advanced AI-based contract analytics, including machine learning-based risk scoring and process automation. Rather than reacting to market and regulatory demands by invoking legacy approaches, they are using the new technologies to substantially reduce AML risk exposure and avoid hefty penalties related to non-compliance, not to mention to minimize the potential reputational blow of such events.
When deficiencies are found using contract analytics, they almost always result from a systematic effort at automated monitoring because the problems are rarely parcel to the due diligence process of setting up the contract or papering, in the first instance. Examiners invariably discover that the level of monitoring does not align to the risk or the type of relationship.
Automation of contract investigation
Companies too often fall into a false sense of complacency that says when due diligence is completed, the work of compliance is mostly done. Traditionally, wrongdoing is almost always identified only when something goes wrong or when a regulator steps in. Like most industries, banking and financial services firms are concerned with the bottom line and the impact of significant cost centers, such as compliance and risk functions.
In an environment of growing regulations, lean staff, and even leaner budgets, there is no way for banks and financial institutions to achieve, let alone sustain, AML compliance without automation. It is simply not possible to extract contract intelligence amid the byzantine world of trade finance with spreadsheets and manual data entry. Management of incumbent contractual details across often hundreds of operational centers and analysis of the massive number of data points needed to satisfy regulatory requirements is tremendously challenging without some form of technological intervention.
Especially amid the rising costs of non-compliance, as fines are levied by regulators with greater frequency and intensity, AI contract analytics has emerged as a gateway for understanding risk and non compliance. As a matter of contract governance and strategy, banks and financial institutions are applying data analytics to their legal agreements for robust oversight of finance, shipping and insurance interests, cross-jurisdictional legal systems and customs procedures, and even in cases where multiple languages govern a relationship.
Automation of contract investigation for AML adherence can also significantly reduce the caseload assigned to manual reviewers. By escalating fewer contracts for investigation by examiners, contract AI is being employed to reduce level-one alerts, populate forms and databases for reporting, and allow for overall labor optimization so that humans can focus on higher-level review requirements. Artificial intelligence also enhances the prioritization of contracts for upstream resolution with rules-driven document management and escalation.
By embedding contract discovery, extraction, and analysis technologies into organic processes for AML compliance, banks and financial services providers can fundamentally reshape their defenses. Keeping pace with threats from organized crime, terrorist elements, and other corrupt cartels and individuals, who often cloak their malicious activities by flying under the contractual radar, requires dynamic approaches that scale to a clear view of risk across an entire organization. With the right application of AI-derived analysis to data contained in contracts and legal paper, these institutions can confidently address the AML challenge.
Stuart Brock is a director at Seal Software where he helps lead Seal’s financial services programs. He is a licensed attorney who practiced law at a top national firm for some 10 years before moving in-house at Bank of America. Stuart’s roles at BofA included oversight for governance and management of third-party contracts including maintenance and technology to support BofA’s arsenal of contract templates across more than 60 countries.