Cookie information
Dear Visitor! We use cookies to secure the site's functionality, improve user experience, analyze website traffic and support our marketing activity. By clicking “Accept Cookies”, you consent to store on your device all the technologies described in our Cookie Policy.
Accept Cookies
Deny Cookies

Streamline your AML with AI

Beyond the hype

The UK’s money laundering problem

£100 billion “Money laundering costs the UK more than £100 billion a year. It is used by criminals and terrorists to move funds and pay for assets.”
National Crime Agency (NCA), 2019
  • It is estimated that less than 1%
    of criminal funds flowing through the financial system worldwide is confiscated.
  • At the same time, most of the banks have been already fined for Anti-Money Laundering (AML) deficiencies. The AML fines for financial institutions in the UK totalled
    £86 million in 2020.



DOWNLOAD OUR SOLUTION BRIEF


The role of the financial industry

Beyond the hype
Know Your Customer (KYC) and Transaction Monitoring are both vital parts of an AML Program.

The main objective of KYC is to verify customer identity and determine the nature of the business relationship.

Transaction Monitoring is a process for detecting suspicious behaviour.
Financial institutions need to make sure that the activity of their customers is consistent with their risk profile and what they know about them and their business.

A typical money laundering scenario

overreporting of revenues using front business

John is making lot of money selling drugs.
He is now sitting on a pile of cash that he has earned with his business.
He is thinking about how he could stash his cash in a safe bank.
He receives a tip to open a restaurant.
He rents the place, pays the owner in cash.
Depositing his money, claiming it as revenue from his restaurant.
Whenever he needs some money to spend, he uses the money on his bank account.
John now looks like a successful restaurant owner making decent income from his legitimate business.

AML challenges for financial institutions

Beyond the hype
Regulatory pressure
Highly manual investigation processess
Immense volume of data
Re-engineered ways to conceal the flow of funds

Financial institutions are facing ever increasing challenges regarding their AML program.
Technology should be there to help. However, traditional rule-based transaction monitoring systems are hammering AML staff with a high number of false-positive alerts.

  • AML
    compliance
    cost
    breakdown
  • human resource cost
    00%
  • IT investment
    00%

AML budgets are not balanced: IT investments are still proportionally low compared to the continuously growing human costs.




DOWNLOAD OUR SOLUTION BRIEF


Artificial intelligence (AI)
and machine learning

Progression
Artificial
Intelligence
Artificial Intelligence
Machine
Learning
Machine Learning
Deep learning
Deep learning
Graph

Why are these becoming widely used?

The roots of these technologies point to the mid of the 90’s.

01 The amount of data available and

02 computation power

were both considerable constraints at the time. The advent of AI and ML coupled with the growing availability of mass data and increase in computational power make deep learning effective today.




DOWNLOAD OUR SOLUTION BRIEF


Applying Machine Learning Methods in AML Transaction Monitoring

Beyond the hype
Supervised
Machine Learning:
Using historical alerts associated with suspicious behaviour true positive alerts can be tagged and a model can run on this dataset to learn how to score and prioritize alerts. Consequently, investigators can focus on alerts that are more likely to contribute to Suspicious Activity Reports (SARs).
proactive alert proactive alert
non-productive alert non-productive alert
traditional scenario cutoff traditional scenario cutoff
ML cerated cutoff ML cerated cutoff
Unsupervised
Machine Learning
AML programs are based on the concept of finding suspicious activity, but there is no objective definition of what suspicious is. Anomaly detection techniques address this issue by identifying outliers; clients and behaviour that appear mathematically distant from the expected.
no-anomalous entity no-anomalous entity
anomalous entity anomalous entity

AI maturity levels

Beyond the hype
complexity of the model
level 01
Defining segments based on historical data to find true positives that were not identified by existing detection methods. The analysis of true positive alerts can uncover hidden connections that go beyond traditional customer segmentation.
level 02
By using Machine Learning tools, the priority of the alerts can be automatically assigned. Based on their risk appetite, one can set objectives regarding what error types are tolerated and the confidence level for with which an alert should be investigated.
level 03
By using AI across the entire Transaction Monitoring process, alerts can be auto-triaged, auto-enriched and narratives for reporting can be auto-generated using Natural Language Processing.

Benefits

Beyond the hype

Regulators must approve and results produced by AI must be explainable.

  • Find out how you can start improving your AML process with AI! Download our solution brief here!

1/7

About

Beyond the hype
Relying on in-depth knowledge and experience our goal at Consortix is to make AML compliance easy for our clients. As a SAS partner, we offer flawless AML project execution and migration to the latest analytics-driven platforms combined with AI /machine learning technology.
Privacy Policy

© 2021 CONSORTIX | Designed & Powered by Positive Adamsky
Streamline your AML with AI