Review Findings Instead of Articles

We are very excited to introduce Findings.

Where others are focused on improving a tedious and error-prone manual screening process, we are focused on eliminating it. This enables risk analysts to focus on value-add decisioning instead of reading long lists of articles.

Ondrej Prostredník

Ondrej Prostredník

September 10th 2020
Share on

Findings are a new way of looking at Adverse Media results. Instead of presenting individual articles, Merlon now shows aggregated findings cards that include information on associated crimes, legal status, and snippets with evidence from multiple related articles. Thanks to the finding cards, analysts don’t need to review the same story over and over, but rather review holistic findings.

Evidence snippets provide analysts with a more detailed understanding of the associated risk while enabling direct access to the underlying full-text articles. Evidence provided in snippets allow you to make a decision at the snippet level without reading the full articles.

Let’s investigate an individual named Jamal T. Hassan who provided material support to Al-Qa’ida and was charged with terrorism offenses. Merlon organized the evidence found in 38 articles into 10 findings which reduced the number of results that must be reviewed by 70%.

In the single finding card below, Merlon presents key evidence from 16 related articles. Without reading the full articles you can promptly conclude that Jamal T. Hassan is under investigation for crimes associated with terrorism and terrorist financing.

Risk Detected

Merlon assigns a RISK DETECTED tag for findings where it identified material relevant risk and the search term has been identified as the perpetrator of the risk.

No Risk Detected

Merlon will also mark some findings as “No Risk Detected”. This happens when:

  • Subject’s name has been detected, but isn’t used as a name
    Merlon has detected search criteria in one or more articles, but the terms are not meant as names. For example, if you search for an individual named Susan Bridge, and the article contains the text “police reported today that the Susan Bridge collapsed due to faulty construction,” Merlon will tag this as “No Risk Detected”.
  • Finding does not contain any relevant risk
    Subject’s name has been detected but no relevant risk has been identified.
  • Subject is not the perpetrator of the identified risk
    Both name and risk have been detected but the investigated person or entity is not responsible for the identified risk. This may occur if the search term refers to a bystander, police officer, witness, or other subject individual or entity mentioned in the article that is not the perpetrator of the risk.

The efficiency gained through Merlon findings can save analysts upwards of 90% on a single investigation. When investigating Lei Guan, former Chinese researcher at the University of California, Merlon organized the evidence found in 112 articles into 54 findings.

Merlon does some additional heavy lifting for you – 51 findings were automatically marked as “No Risk Detected”. In this example among the dismissed findings there are mostly articles related to patent applications or scientific papers which do not contain AML/KYC risk.

Therefore in this case the analysts would only need to review  3 well selected findings to understand that Lei Guan was charged with destroying evidence related to illegal transfer of US tech.

What else is new?

Merlon has made a number of other improvements to your experience:

  • You can now see if an investigation contains risk or no risk right from your investigations list:
  • A findings summary box now tells you how many articles have been grouped into findings as well as how many had no risk detected:
  • It’s now faster to sort by Relevance or Date thanks to dedicated sort buttons. You can change the sort order by clicking on the appropriate button twice:

We built Merlon to allow you to instantly use the power of AI to fight financial crime.