What is screened
Customers are screened at onboarding and whenever the lists change, and payments, particularly wires and international transfers, before processing. The lists include OFAC's Specially Designated Nationals list and, depending on jurisdiction, those of the UN, EU and UK.
Why there are so many false positives
Names are matched loosely to catch transliterations, aliases and misspellings, so common names produce many false positives. Each hit must be reviewed and the reason for discounting it recorded, and a true match can mean blocking or rejecting a transaction and reporting it.
Where the work is
Reviewers compare the listed party's identifiers, such as date of birth, nationality and address, with the customer's and read payment messages for context. The work is repetitive, but a wrong decision has severe consequences, so it is still usually done by hand.
Where automation helps
An AI system can compare identifiers, read payment messages and adverse media for context and clear hits that differ on decisive identifiers, with a written rationale, once its accuracy has been measured and the sanctions team agrees. Every possible match goes to a person, and the screening engine and thresholds stay unchanged.