Go Back

Artificial Intelligence: Competitive Advantage for Financial Institutions The Regulatory Landscape

AI and Compliance

Financial Institutions evidence compliance with regulatory rules by embedding them into their policies, procedures, and controls. The traditional approach to updating and managing these changes is very manual and partially consultancy led/dependent. Technologies like Artificial Intelligence are being harnessed instead to transform their approach to governance, compliance and risk.

Artificial Intelligence (AI) is an umbrella term for a series of underlying technologies such as Machine Learning (ML) and Natural Language Processing (NLP), that can be brought together within a cloud-based environment to store and process huge amounts of data, to perform sophisticated tasks, without the assistance of humans. ML uses statistics to find patterns data which are used to make predictions or perform tasks. Natural Language Processing, refers to the interactions between computers and human languages. The key objective of NLP is to read, decipher, understand, and make sense of the human language so that it offers a valuable output.

Question Answering with AI allows you to identify targeted queries to large amounts of text quickly

Applications of Artificial Intelligence

Below are some examples of AI and its application to a compliance function.

Semantic similarity is an NLP task that helps evaluate the semantic distance or proximity between texts. In layman’s terms, this means finding similarities between content. An example of this would be automating the mapping of related content together, identifying relationships between a specific regulation and relevant related policies, procedures, and controls. This gives compliance professionals an instant impact analysis of whether internal governance documentation is in line with related regulatory obligations or their risk framework. A second example could be identifying similarities across content so that you can consolidate documents that repeat one another. This can consolidate governance documents by up to 50%.

Contradiction and inconsistencies, NLP functionality that applies a single scale identifying cases from ”inconsistent” to “consistent”. A common use for this algorithm is identifying contradictions between an organisation internal governance documentation. Typically policies and procedures are written and/or updated in silos across a business leading to contradictions. Identifying these contradictions can help achieve a unified and standardised governance framework.

Question Answering allows you to identify targeted queries to large amounts of text quickly. The more text you have in your system, the smarter this feature becomes. Smart applications of this include implementing a chatbot on an employee/customer portal to help provide quick answers to technical questions or unfamiliar regulatory guidelines.

Paraphrase Detections goal is to define cases when one phrase was expressed using another – ‘rephrase’ cases. When financial institutions write their standards and policies, very often they simply paraphrase or rephrase original paragraphs into similar text with the same meaning. Similar rephrased content scattered across policy documents can have synonyms and different language structures. This functionality allows compliance officers to automatically identify paragraphs with similar interpretations, connecting them to previously prescribed clauses in other documents. In addition, when a new regulation is enacted with slightly different content, it is possible to map it to previous versions. Therefore, a base of related paragraphs linked with regulations is built up which can then be automatically linked to other relevant content.

Named Entity Recognition. The ML functionality analyses words that have different endings so that they can identify and group documents and paragraphs into themes. It detects that various combinations of symbols still refer to the same entity. Just like “NY” and “New York”. It is helpful for search and also useful for statistical-based reports. For instance, to be able to evaluate fast, what investment types have been covered by new regulations and the frequency across documents. For example, multiple mentions of insurance-related investment types will provide insight to the increase in the risk for that type of product at the bank.

AI and the Future of Compliance

Compliance is evolving at a rapid pace. The technology over the course of the next 10 years, AI-based advancements will radically shift it from what it looks like today. AI will assist compliance professionals by automating time-consuming administrative tasks, whilst enhancing decision making. It will prompt them with the right decisions, highlight potential errors, and allow them to carry out tasks more efficiently and accurately.

This will, of course, require, new systems and more advanced tech skills from compliance professionals. A large portion of FinTech and Challengers have this in place due to their technology-focused hiring and lack of legacy systems. The global banking sector is also following suit. Two key themes across the 2018 annual reports were around upskilling their workforce and using technologies like AI to improve operational efficiency.


Follow Us

Subscribe to Shield’s Newsletter

Capture everything. Deploy anywhere. Store in one place.