How does fintech help fight market abuse before it happens?


When it comes to detecting and stopping misconduct or non-compliant behavior, monitoring teams must act like detectives – curating large volumes of data, uncovering the truth and acting on what They find. Teams have the difficult task of essentially finding the needle in the haystack, just to make sure the few cases of market abuse are stopped and reported.

Yet, despite the fact that proactive email monitoring has been a regulatory requirement for some time, many organizations rely on outdated monitoring methods. These approaches depend on the use of certain search terms or lexicons to identify potential issues and alert compliance teams. This legacy approach to fault detection causes a flood of mostly irrelevant alerts – sometimes hundreds of thousands a day – that need to be addressed by examiners.

The word problem

Simply flagging risk words is not an infallible system for detecting mistakes. For example, in everyday conversation, I can tell someone that I guarantee they will have a good time at the football game or party this weekend. But if I’m a financial adviser communicating with a client, I can’t say I’m guaranteeing something – especially a certain stock that goes up tomorrow because it’s a risky asset and there’s a good chance that she drops.

So when a legacy surveillance system monitoring email sees the word “guarantee” in a message or other communication, it flags the message for review without understanding the context. The problem is that we use common words like collateral every day in other contexts that have nothing to do with compliance risk or financial markets. When you rely solely on lexicons, you’re mostly looking at unwanted communications that have nothing to do with the risks you’re looking for.

A factory on the alert

The main objective, when we talk about monitoring communications in the financial sector, is to prevent problems before they become real problems. We hear a lot about misconduct – which can refer to things like money laundering, collusion or insider trading – in financial services, and many of them could have been detected quickly with proper monitoring. proactive. But to do this, monitoring platforms need to drastically reduce false positives in order to have the bandwidth to investigate meaningful risks.

Historically, surveillance has rather been a production of alerts. Organizations have an army of reviewers who spend their days reviewing a mountain of alerts, mostly irrelevant ones. They review, climb and rehearse day after day, mostly sifting through trash. When you have so many false positives, an army of people is wasting a lot of time reviewing and clearing those alerts.

A better approach

It starts with two technology areas that are booming today: artificial intelligence (AI) and machine learning. The goal is to make all communications including email, chat and audio searchable, leveraging the power of AI and machine learning to present only the most relevant data. to examiners.

Modern communications monitoring platforms seamlessly extract communications from every enterprise communications system to ensure you have comprehensive coverage. An intuitive user interface with advanced workflow configurability, adapts to your organizational workflows securely in the cloud. Machine learning always works in the background and gets smarter with each review to deliver only the riskiest content and reduce false positive alert volumes.

When an alert is valid and calls for a real investigation, examiners can use the same modern system to initiate that investigation with built-in case management.

Another key feature is email thread deduplication – which locates old email content in new emails, verifies that it has already been scanned, and prevents it from generating another alert. The data cleansing capabilities of modern systems identify email headers, signatures, and disclaimers. They prevent the generation of alerts on this duplicate or uncreated content, so that reviewers are only informed of the relevant part written by the sender. Reviewers can also see what content has been erased, providing full transparency into how machine learning is working.

With all of the great advancements in surveillance and modern systems available, the obvious question is, why haven’t all organizations transitioned to this future state? One answer may be the historical lack of transparency and explainability of these models.

Lexicons, although ineffective, are easily explained (for example, this e-mail was alerted because it contained this word). There are a lot of fears in the industry about the transition to machine learning models because now you have to defend it. As we move forward, it behooves surveillance innovators to ensure that all AI is transparent and explainable, and ultimately defensible to regulators. Tech companies will continue to invest in these new technologies to help make processes more accurate and efficient and help monitoring teams catch faults. Meanwhile, users will be able to uncover the truth in the data and act accordingly to combat manipulation.

Jordan Domash is Relativitygeneral manager of Relativity Trace. He guides a focused team in the development of this tool, built on the Relativity platform, for proactive compliance monitoring, supporting engineering, marketing and sales.

Previous BDI introduces the Interval collection of home entertainment cabinets and consoles
Next Obituary: Pierre Zakrzewski, Irish cameraman killed in Ukraine who brilliantly covered overseas conflicts