Risk and Compliance is no longer the black and white, tick-the-box function it once was. Today, compliance teams are faced with increasingly complex, multi-dimensional challenges that continue to change year after year.
With new regulations being introduced by the FCA and on a regular basis, it is becoming more difficult for financial institutions to satisfy and stay on top of their obligations. In fact, recent research from LexisNexis Risk Solutions suggests that over half of regulated firms in the UK are risking penalties for non-compliance with the Fifth Money Laundering Directive (5MLD), legislation introduced in early 2020 to combat the growing money laundering problem facing the UK (Financial Reporter). GDPR compliance is also a growing issue, with the 12-month period from July 2019 to June 2020 seeing a 260% increase in monthly GDPR fines compared to the same period the year before.
Risk and Compliance is growing in its complexity, and Fintechs and incumbents alike are feeling the impact. Many are looking to Artificial Intelligence (AI) as a potential solution, and with risk management representing the leading AI implementation area among financial service domains, great strides have already been made. Under this umbrella of AI, Natural Language Processing (NLP) is among the fastest growing branches, and many compliance teams are looking to use NLP to transform the way they operate.
What is NLP?
While the field of Natural Language Processing has its roots in the 1950s, the past ten years have seen NLP really accelerate to become a part of our daily lives. From Alexa to autocorrect, you likely interact with some NLP-enabled device or technology on a weekly, if not daily, basis. Broadly speaking, NLP has seven key capabilities, which are:
- Sentiment Analysis
- Topic modelling
- Text categorization
- Text clustering
- Information extraction
- Named entity resolution
- Relationship extraction
Within the Risk and Compliance Functions, these capabilities serve as powerful tools.
Applications of Natural Language Processing (NLP) in Compliance
Turning Unstructured Text into Actionable Insights
One of NLP’s greatest strengths is its ability to deal with unstructured data. If you ask your head of compliance how much data they were dealing with twenty years ago, their answer would probably not be “too much.” However, in the age of big data, having too much information is a very real problem. Today, compliance teams deal with a never-ending inflow of unstructured data: company documents, social media, websites, mail, chats, and voice calls… the list is never-ending. While all these sources contain potentially crucial information, it would simply be impossible for a human to parse this data. This is where Natural Language Processing comes in. NLP can extract data from text and provide it with structure. Then, it can enrich this data through topic categorisation and sentiment analysis, and finally, it can generate actionable insights through Natural Language Generation (NLG).
Already, compliance teams across their UK have been rolling out NLP solutions to conduct due diligence, screen for Anti Money Laundering (AML), and ultimately free up staff to work on more important, higher-value areas. Furthermore, NLP not only empowers compliance teams to deal with data more effectively, but it also enables a more human-centric interaction with customers.
Improving Customer Interaction
While it is easy to think of AI as a step away from human interaction, NLP can actually enhance these interactions. One rapidly growing use case of NLP in Compliance roles that illustrates this is voice technology. With the pandemic forcing financial institutions to adapt to serving clients remotely, NLP has been crucial to helping monitor risk and compliance in this new setting. With customer interactions going virtual, some banks have seen call volume increase as much as 400% since the start of the pandemic (BPFI).
Voice technology offers a solution to dealing with high call volume by transforming unstructured voice data into text and enabling risk assessors to identify and flag potential compliance issues in near real-time. Voice technology also enables companies to transcribe, store, and organise these calls for review, providing compliance teams with a system that allows them to focus on the customer while satisfying regulatory requirements. At Aveni, NLP and voice technology are harnessed to monitor every customer interaction, which empowers compliance teams to identify customer vulnerability, evaluate the conduct and competency of agents and advisors, and ultimately move from a detective to a preventative control framework.
As discussed in a recent Financier Worldwide Article, voice technology can significantly reduce risks to financial institutions. Using voice technology “ensures that historical archives of voice data are transcribed for analysis” and “enriches interactions” by analysing customers’ language and offering prompts and supplementary information on the spot. Overall, utilization of voice technology ensures a compliance system that “gives FIs a better understanding of their customer” and “aids in mapping the customer journey,” all while ensuring regulations obligations are met.
Challenges with NLP
NLP clearly has the potential to transform risk and compliance functions across the financial services industry, but how can companies capitalise on this transition? As discussed in our article How AI is transforming Risk and Compliance, implementing AI effectively is easier said than done.
The most notable challenge for most firms will be acquiring the necessary technical expertise to roll out NLP solutions effectively. While some compliance teams will look to hire those with expertise in the fields of AI and computer science, larger, established firms may find such a restructuring challenge, and may instead look to Fintechs as their source of personalised AI solutions.
Regardless of the approach to adoption, NLP will be a key capability for many compliance teams. For many, it will be the key that allows a transition from the post-hoc investigation approach to compliance with the preventative, predictive approach, materially reducing exposure to risk.