The 1H 2022 Screening Backend Upgrade includes a host of improvements and fixes to the Alessa sanctions screening engine.
Improved name matching algorithm.
There are many improvements that were made to the algorithm, including improved performance, support for numeric digit to text matching, AI model improvements, and bug fixes. The changes made have been tested and bring better name matching output by reducing false positives and increasing accuracy.
Faster API performance.
This release includes a number of improvements in how requests sent to the Alessa API are processed. In our testing this has lead to improved response times for requests sent to the Alessa API for realtime transaction monitoring. No changes are required for your organization to take advantage of these enhancements.
Improved API resilience.
We have improved scaling of our Sanctions Screening Backend which should not only lead to faster performance but improved resiliency, error handling, and less downtime.
Prioritization of ASCII character names vs Native names.
This will prioritize an ASCII name match over Native if they have the same score. This is only relevant to your organization if you are using Native matches.
Keyword selection for Dow Jones research data.
For clients leveraging Dow Jones data you are now able to enable keyword list selection directly within Alessa.
Sanctions screening defects.
As part of this release the following defects have been resolved:
- Screening Configuration - Changes in 5.5.3.19.4 do not allow keyword list changes to be saved/updated
- Screening - Backwards compatibility of screening configuration page
- Screening - Multi Token Names not bringing incorrect results - some names with many tokens will yield a lower than expected score. (eg. Slaheddine Ben Haj Hamda Ben Haj Hassen Ben Ali)
- Realtime transaction performance improvements - entities are not sent as batch for screening