Semantic Technology Transforms Negative News Monitoring Procedures 

Background

One of the most critical tasks that a bank has to undertake is carrying out a thorough investigation of their potential customers. According to the Anti-Money Laundering laws, banks are obliged to apply robust legal procedures to prevent financial crimes and other fraudulent activities.

To comply with these legislative regulations, the financial organizations perform internal procedures called KYC (Know Your Customer), which include comprehensive background checks for each new customer. The ultimate goal is to identify who they are, verify their credentials, and evaluate potential risks. 

KYC is an integral part of another process – Customer Due Diligence. Diligent governance requires global investment banks to use extensive KYC processes. According to the established practice, bank analysts have to scan hundreds of news articles for negative information about their potential customers. They aim to verify whether a news article is related to the ultimate beneficial owner under investigation and, in case of a positive match, to present reliable evidence for negative associations. Such an examination is a daunting, time-consuming task, and it is not very efficient.

The Goal

Global financial institutions have to streamline the process of identifying relevant negative news related to a person or entity of interest.

The Challenge

One major problem of the negative news monitoring process is that a significant number of news articles delivered by third party databases are about an incorrect beneficial owner with the same or similar name.

Another one is that not all parties mentioned in a news article are involved in the negative event (false positives). Therefore, financial organizations need an automated solution that would enable their analysts to work with a selection of news that does not contain irrelevant entities or false positives.

The Solution

Here is where semantic technology can help with automating the process and optimizing the workflow across media analysis. Sirma AI’s smart solution leverages big knowledge graphs built from the proprietary commercial database used by a particular financial organization and enriched with publicly available resources like OpenCorporates and CrunchBase.

It classifies the identified sentences within each news that contain relations between the person/entity of interest and an adverse event. Then, it filters out irrelevant news and scores the rest by relevance and confidence. As a result, bank analysts can receive highlighted sentences with the relevant piece of content, ready for a quick review.

The result

This technology solution is fully automated and enables financial institutions to significantly reduce the time for such analysis to less than an hour. Also, it ensures that banks comply with the regulatory requirements and are protected from reputational damages.

Read more about technology application here.