Every 48 hours, the digital world produces an astounding five quintillion bytes of data, according to McKinsey & Company.
Most of this data is unstructured, presenting significant challenges for financial institutions that typically handle only structured data during anti-money laundering (AML) and KYC operations.
According to Saifr, this limitation can have severe real-world consequences, as evidenced by a 2017 case in which a major global bank was fined £163m by the U.K.’s Financial Conduct Authority (FCA). The FCA’s penalty was due, in part, to inadequate customer due diligence which left the U.K. financial system vulnerable to crime.
This situation underscores the shortcomings of many financial institutions’ Know Your Customer (KYC) programs, the RegTech firm stated. For instance, the bank in question failed to prevent mirror trades that channeled billions of dollars to overseas accounts, a situation that could have been mitigated with a more robust AML/KYC strategy encompassing both structured and unstructured data.
Structured data is neatly organized and formatted, making it easily searchable by conventional software applications through simple queries or keyword searches. Financial entities might use this data type for scanning customer information, transaction records, and other regulated lists as part of their AML/KYC protocols.
Saifr also outlined that conversely, unstructured data includes various formats like emails, social media posts, and videos, which are often messy and difficult to search. This data type is crucial for adverse media screening but is frequently overwhelming due to its sheer volume and complexity.
Currently, many financial organizations rely solely on tools designed for structured data, neglecting the vast majority of unstructured data available online. This oversight leads to inefficient searches and a high rate of false positives. Advanced profile-based screening tools, while useful, still struggle unless they can contextualize the data they analyse.
However, the tide is turning with the adoption of advanced artificial intelligence (AI), machine learning (ML), and large language models (LLM) in the financial sector. These technologies are revolutionizing how organizations handle both types of data, enabling them to perform real-time, comprehensive analyses that capture the context and relevance of the information. For example, AI can discern the sentiment of a news article, helping firms better identify potential risks.
As the digital landscape grows, so too does the sophistication of potential threats. Financial institutions that leverage AI are finding that they can enhance their operational efficiency and accuracy, thereby improving their fraud and risk management capabilities. By integrating AI tools, firms are not only able to detect risks more swiftly but also help ensure comprehensive monitoring across different languages and global online platforms.
Before integrating new screening tools, it’s crucial for financial teams to evaluate their current processes. Questions such as unstructured data, the ability to data in real time, the capacity to handle multiple languages, and the effectiveness of current AI tools in detecting suspicious activities should be addressed to close any gaps in existing systems.
Copyright © 2024 RegTech Analyst
Copyright © 2018 RegTech Analyst
