Skip to content

Screening adjudication

Screening providers return candidate matches. A candidate is not a finding until a human says so, and Verity will not let a review be approved while any blocking candidate is unadjudicated.

What a match is

A hit against a sanctions, PEP or adverse-media list, carrying the matched name, the list or watchlist, and the provider's classification. It is evidence for a decision, not a verdict — the screening panel says so.

Adjudicating

Needs screening.match.adjudicate. For each match, record whether it is:

  • A true match — this really is the person or business on the list
  • A false positive — a name collision or otherwise not our subject

Adjudication is audited, and a false-positive decision emits screening_match.false_positive.

Blocking versus non-blocking

Which matches block approval depends on the screening disposition configured for the entity's country — see Country defaults. A country may be configured so that a PEP match blocks, or escalates, or merely informs.

A blocking match that has not been adjudicated stops approval outright. This is the only hard block in the risk surface; the gates are advisory by comparison.

Reverting

An adjudication can be reverted, and doing so re-blocks approval in place. This emits screening_match.adjudication_reverted. Voiding an adjudication emits screening_match.adjudication_voided.

Reverting exists because adjudication is a judgement made on the evidence available at the time, and new evidence can change it. The audit trail keeps both decisions.

Pending results

A screening that has been requested but not answered has no match count — not a count of zero. The panel says the result is pending. Zero would mean we looked and found none, which is a different claim entirely.

PEP on approval

Approving a review while a PEP match is present is permitted where the country's disposition allows it, but the approval is stamped with the PEP fact and the matched names. The record shows the decision was informed rather than accidental.