OFAC Screening Threshold

The similarity score at which a potential name match is flagged for review. Lower thresholds catch more but generate more false positives to review.

TL;DR

TL;DR: A screening threshold is the match score above which a potential name hit is flagged for human review. Lower thresholds catch more true matches but generate more false positives.

How a threshold works

Screening engines compare a counterparty name against the SDN List and produce a similarity score. The threshold is the cutoff: names scoring at or above it are flagged for review, and names below it are treated as clean. A high threshold means only near-exact matches are flagged. A low threshold casts a wider net and flags fuzzy and partial matches too.

The tradeoff

Setting the threshold too high risks false negatives, where a real match scores just below the cutoff and slips through unnoticed. Setting it too low floods reviewers with false positives and invites alert fatigue, where real matches are missed because reviewers stop paying attention. The right setting balances catching listed names against a review queue a team can actually clear.

Scoring helps more than a single line

A single cutoff treats every flag equally. A risk score adds context, ranking flagged names so reviewers address the most concerning ones first. tools with a risk_score capability, such as sanctionsai.dev, let teams set a threshold for the queue while still surfacing the highest risk items at the top.

Why it matters

Because OFAC liability is strict, a missed true positive is a violation even if the tool said clean. A well-tuned threshold, combined with scoring and current list data, is the difference between a screening system that protects the organization and one that only looks like it does.