A false positive is a match that, on review, turns out not to relate to the person or entity searched for. Reducing the number of alerts does not by itself demonstrate control quality.
Start with data quality
The consistency of the name field, alternative spellings and additional identity details affects the assessment. Do not compensate for missing information by assigning a higher confidence score.
Thresholds and samples
Do not measure threshold changes only by the number of closed alerts. Also assess the risk of missing true matches and review samples that represent different name groups.
Preserve decision learning
Record review rationales consistently. Define the conditions under which a previously closed result is reopened when source or identity information changes. Do not promise a universal false-positive reduction percentage.
See the related product scope → AML screening
Legichain — technology for identity verification, AML and blockchain screening. Contact: [email protected]