How to reduce PEP & sanctions screening false positives
A screening programme that fails every namesake is as broken as one that clears everyone: analysts burn out on noise and real risk hides in the pile. The fix isn’t a looser or a tighter dial — it’s matching on more than a name, and putting a person on the ambiguous cases.
Five practical levers
- Use secondary identifiers — date of birth, nationality and other attributes turn a name “maybe” into a confirm or a reject.
- Tune thresholds per list — a sanctions list and an adverse-media feed don’t deserve the same score cutoff; set them independently.
- Whitelist cleared entities — once a reviewer clears a specific false match, remember it so it doesn’t re-alert on every re-screen.
- Categorise, don’t dump — surface why something matched (which list, which risk type) so a reviewer decides in seconds.
- Keep a human in the loop — auto-clearing to cut noise is how real hits slip through; route the ambiguous ones to review instead.
Why maker-checker matters
The cheapest way to make an alert queue small is to auto-clear low-confidence matches — and it’s also how a real sanctions hit gets missed. A maker-checker (four-eyes) workflow lets one reviewer decide and another confirm, so you get precision and an auditable record of who cleared what and why. Zanyara builds this in, and applies the same triage to onboarding screening and to ongoing monitoring alerts.
The balance to holdPrecision (fewer false positives) and recall (no missed true hits) trade off against each other. Good tooling doesn’t pick one — it moves the ambiguous middle to people, and automates only the clear ends.
Frequently asked questions
- Why are PEP false positives so common?
- PEP and sanctions lists match on names, and names are ambiguous. A common name, a transliteration, or a partial date of birth can all produce a “match” against someone the customer isn’t. Screen a large book with loose settings and most of your alerts will be namesakes.
- Won't tightening thresholds cause missed hits?
- It can — which is why you don’t just tighten. The goal is precision without losing recall: use secondary identifiers to confirm or reject rather than only raising the score cutoff, and send the genuinely ambiguous cases to a human instead of auto-clearing them.
- How does Zanyara help?
- Zanyara combines tunable matching, secondary-identifier checks and whitelisting with a maker-checker review workflow, so cleared entities stay cleared and every decision has an auditable owner. The same triage runs at onboarding and under ongoing monitoring.