Screening: Reduced false positives in Jaro-Winkler matching
Screening matches using the Jaro-Winkler fuzzy setting are now scored per name token instead of across the whole name at once, and weighted by token length. This significantly cuts down false positive hits without reducing detection of genuine matches, including transliteration variants and partial name matches.
You can use this while setting up screening rules shown in screenshot below:
Custom rules: String length function
You can now measure the length of a string in a rule condition and compare it against a number. This works in both custom transaction rules and custom screening rules. For example, you can require a screening match to be longer than 3 characters before it counts as a hit.
Key updates include:
New
String lengthfunction, usable anywhere a numeric comparison is available.
Console: Copy a risk factor logic configuration
You can now duplicate an existing risk factor logic configuration instead of rebuilding it from scratch. The copy is inserted directly below the original with "(copy)" added to its name, and is fully independent, so editing one does not affect the other.
Key updates include:
Copy action available on risk factor logic configurations in edit mode.
Also available on list variables.
Screening: Select "Profile of interest" without a PEP tier
"Profile of interest" can now be selected as a screening source on its own. Previously, at least one PEP tier had to be selected alongside it, even when Profile of interest was the only relevant source.
Case management: Hit reason for alerts and screening
You can see exactly why an alert or screening hit was triggered, instead of having to reverse-engineer it manually. This is an existing feature we recently added a knowledge base article for available here:
