Your single source for new lessons on legal technology, e-discovery, and the people innovating behind the scenes.

The Case for Al in Communications Surveillance

Jennie Brotherston – GreySpark Partners

Surveillance is high on the agenda for most financial services’ compliance teams. Despite extensive work done in this space, many firms still find their surveillance processes insufficient compared to the expectations of regulators.

Earlier this year, the UK’s Financial Conduct Authority (FCA) issued a statement that it remains concerned that requirements for market abuse surveillance are not being fully met, despite the Market Abuse Regulation (MAR) having been introduced as far back as 2016.

The reason that market participants are not fully meeting all requirements may not be for want of trying, however. Surveillance—and communications surveillance, in particular—is a complex and data-heavy activity. Consequently, many compliance teams are now looking at how artificial intelligence (AI)—or one or more of its subcategories of advanced analytics—could be applied to the puzzle, either alongside or in place of more traditional solutions.

Indeed, UK, US, and Canadian regulators, amongst others, are themselves focused on investing in advanced technologies, such as AI, to detect and prosecute firms participating in market manipulative behaviours. For instance, the Monetary Authority of Singapore (MAS) is using predictive modelling to identify individual financial advisors that have a higher likelihood of being involved in market misconduct. Mirroring regulators, compliance professionals across the globe are reviewing their legacy surveillance systems and looking to implement “next generation” surveillance solutions. A 2021 survey conducted by GreySpark Partners of key compliance personnel in 18 capital markets firms operating across EMEA, the US, and APAC indicated that as many as 70 percent were looking to future-proof their surveillance solutions with advanced analytics.

The majority of next generation surveillance solutions available today use AI techniques to process, manage, and analyse the large, complex, and diverse volumes of data that compliance teams need to monitor. As firms begin to embrace AI in pockets across their organisations, many people may still have questions about what the application of AI technology means in practice. A new report by GreySpark Partners, in association with Relativity, presents an explanation of artificial intelligence and describes why it is a particularly useful set of technologies to apply to the communications surveillance use case.

Problem #1: There’s too much data to review comprehensively.

Today’s data volumes are immense. Compliance teams need to be able quickly sort out the irrelevant material and focus in on what really matters.

Related Challenge: Filtering Irrelevant Content

Items that appear in surveillance data sets but are irrelevant include internal newsletters, blog posts, marketing materials, and other mass messaging, as well as non-authored content like email headers and signatures, disclaimers, and confidentiality statements.

How AI Can Help:

  • A good filter will rank content for relevance and automatically remove materials that don’t meet the appropriate threshold.
  • A machine learning-enabled filter will, over time, optimize irrelevant content removal based on the specific needs and standards of an organisation and their compliance requirements.

Related Challenge: Duplicated Content & Email Branches

Communications data sets often contain a lot of duplicated content, such as copies of the entire historical conversation in emails. Conversational context is important, but moving through the same data over and over is inefficient.

How AI Can Help:

  • Intelligent email threading allows emails to be grouped together so that the whole conversation can be identified and any applicable alert is sent only once.
  • A good machine learning-enabled threading and deduplication tool can recognise linear conversations and identify conversations that branch, as well as added participants, attachments, and content. As a result, alert volumes may be reduced by up to 60 percent.

Problem #2: Data comes in diverse formats.

Communications surveillance isn't only about reviewing emails, chats, and text messages. The presence of audio recordings and other formats requires technology that can analyse these data types with equal efficiency and effectiveness as it can written conversations.

Related Challenge: Voice Data

The variety of jargon, acronyms, slang, languages, and codes used in communications can make achieving high­quality voice transcription and interpretation challenging.

How AI Can Help:

  • Effective surveillance processes must normalise voice and text data before analysis.
  • An intelligent NLP-enabled transcription tool can recognise words spoken in many languages, dialects, and accents, as well as conversations in a mix of languages and those that include idiomatic speech patterns, and normalise voice data into a searchable text format.

Problem #3: There are too many false positive alerts.

With so much data to manage, traditional surveillance platforms often trigger an increasing number of false positive alerts for compliance teams to review. This is a major hurdle to their productivity and introduces room for human error as they are inundated with mostly irrelevant data to decipher.

Related Challenge: Risk Evaluation

Risk ranking is the backbone of any surveillance process, but it would be prohibitively time consuming to do it manually.

How AI Can Help:

  • An effective surveillance solution ranks each item according to a risk scale, so that highly ranked items can be reviewed manually.
  • A machine learning-enabled process learns from the human review of each alert and can quickly reduce the volume of alerts.

Related Challenge: Pattern Identification

It is rarely possible for the same person to manually review all content and, therefore, detect patterns of behavior across a wide data set.

How AI Can Help:

  • The identification of unusual or unexpected behaviours can facilitate detection of anomalies and help identify truly risky behaviour.
  • Machine learning techniques can identify patterns in the data set, including, for example, behavioural trends.

Related Challenge: Advanced Lexical Searches

Simple lexical searches generate many false positives and lexicon-based solutions quickly become outdated.

How AI Can Help:

  • An effective lexicon-based system will do more than simple word matching and will be continually updated as risks and parlance evolve.
  • Al can intelligently search text, adjusting for mistakes, jargon, and context, and Al analysis tools can assist compliance teams in keeping lexicons and associated rules up to date.

Problem #4: Communication happens across many platforms and media types.

As the normal course of business communications now moves across diverse applications, including chat, video, and audio, compliance teams are left with a complicated web of data and conversations from across the corporate environment.

Related Challenge: Channel Switching

Without being able to identify pieces of conversation that switch between different channels, reconstructing conversations is challenging.

How AI Can Help:

  • An effective surveillance solution will be able to identify and track all elements of a conversation between a given group of people across all media, and then analyse the content of that conversation.
  • Al categorisation and visualisation capabilities can sort conversations into subsets of data based on timing, participants, concepts, or topics, to identify unusual content and facilitate a more accurate ranking.

AI is the Future of Compliance Best Practices

Surveillance is inherently data driven, and the volume, complexity, and diversity of data can be extremely high—particularly for unstructured, non-numerical communications. This leaves traditional database and analysis technologies struggling to manage, query, and analyse efficiently.

An effective surveillance system should be able to identify only interactions which are truly risky, so that the human reviewer is able to quickly home in on real problems. However, machine learning and natural language processing could offer surveillance teams a next generation solution to this problem.

The Case for Artificial Intelligence in Communications Surveillance

Jennie Brotherston is a consultant at GreySpark Partners, a capital markets consulting company serving investment banks, hedge funds, and asset management firms.