Sanctions and PEP Screening

Beta. The screening section and its results may change without notice. It is a first reading to help your team review faster, not a compliance decision: a person on your side must review every possible match. Tell us what you need at support@topograph.co.

Every company response now screens the physical persons it returns against sanctions and politically exposed person (PEP) lists. You do not request it: when a response carries legal representatives, beneficial owners, shareholders or other key persons, it also carries a screening section. Each person gets one result: no match, a possible match to review, or matches that were all ruled out with a reason. During the beta, screening is not billed. All examples on this page use fictitious data.

What is screened

Persons. Every physical person (up to 200 per company) in legalRepresentatives (including the person who represents a company director), ultimateBeneficialOwners, shareholders and otherKeyPersons. A person who appears in several of these is screened once, and appearsIn lists where. The company itself and its corporate shareholders are not screened yet. Lists. Topograph’s own collection of sanctions lists and PEP lists, read from the official publishers: the UN Security Council, the EU, the UK, the US Treasury, Switzerland, France and other national lists, plus parliaments, governments and other official sources of politically exposed persons. Matches can also come from lists of close associates of PEPs. Identity. Names are matched together with the birth date, nationality and country of residence the register gives, when it gives them. A birth date makes the biggest difference: it rules out most namesakes.

How a match is read

Every match carries an assessment, an automated first reading of whether the register person and the list entry are the same person.

Birth dates are compared by a fixed rule. Birth years two or more years apart make the match different_person; a one-year difference is never enough. The same birth month, with compatible names, makes it likely_same_person. Both are decided by rule, with the dates compared in reason.
An AI reading looks at the rest: names, aliases and positions. It can only rule a match out when the names are clearly different, not mere variants (another order, a missing middle name, a transliteration, an initial). It returns a verdict without a written reason.
A sanctions match is never ruled out by the AI reading alone. At most it stays undetermined, and a person must look.
Missing data is never evidence. When nothing tells the two apart, the match stays undetermined.

The person’s result follows from their matches: ruled_out only when every match is different_person, possible_match as soon as one is not.

When it arrives

Screening runs as soon as the persons are known, alongside the rest of the request. Until it finishes, the datapoints that carry persons show enriching in request.dataStatus, then turn succeeded together with the finished screening. The request still reaches its final state once, with one final webhook, as it always has. Screening usually takes about a second. If it cannot finish in time, the request completes with screening.status set to failed. Screening never fails the rest of the request.

Response

{
  "screening": {
    "stage": "beta",
    "status": "completed",
    "screenedAt": "2026-11-03T10:15:00.000Z",
    "categories": ["sanction", "pep", "close_associate"],
    "persons": [
      {
        "entityId": "person_001",
        "name": "Ivan Petrov",
        "appearsIn": ["ultimateBeneficialOwners", "shareholders"],
        "result": "possible_match",
        "matches": [
          {
            "entryId": "eu-fsf-eu-1234-56",
            "name": "Ivan PETROV",
            "categories": ["sanction"],
            "sources": [
              { "id": "eu_fsf", "title": "EU Financial Sanctions Files (FSF)" },
              { "id": "gb_fcdo_sanctions", "title": "UK FCDO Sanctions List" }
            ],
            "score": 0.94,
            "birthDates": ["1961-03-02"],
            "countries": ["ru"],
            "assessment": {
              "verdict": "likely_same_person",
              "decidedBy": "rule",
              "reason": "Birth dates agree: 1961-03 in the register, 1961-03-02 on the list."
            }
          }
        ]
      },
      {
        "entityId": "person_002",
        "name": "Jeanne Martin",
        "appearsIn": ["legalRepresentatives"],
        "result": "no_match",
        "matches": []
      }
    ]
  }
}

Fields

Field	Description
`stage`	Always `beta` for now.
`status`	`pending` while screening runs, then `completed`, or `failed` when it could not run.
`screenedAt`	When the lists were checked.
`persons[].entityId`	The person’s `entityId` in the same response, to find them in the other sections.
`persons[].appearsIn`	The sections the person appears in.
`persons[].result`	`no_match`, `possible_match` or `ruled_out`, over every match found.
`persons[].matches`	The five most relevant matches, those still standing first.
`matches[].categories`	`sanction`, `pep` or `close_associate`.
`matches[].sources`	Every list that carries the entry. The same person listed by several publishers is shown once.
`matches[].score`	How closely the names match, from 0 to 1.
`matches[].assessment.verdict`	`different_person`, `undetermined` or `likely_same_person`.
`matches[].assessment.decidedBy`	`rule` (birth dates), `model` (the AI reading) or `none` (not assessed).
`matches[].assessment.reason`	The values compared, when a rule decided.

Where screening is not included

Prefill mode. Requests in prefill mode (formerly onboarding) are not screened. Screening runs in verification mode, the default.
Development environments. The sandbox does not screen. Its responses have no screening section.
Monitoring notifications. Monitors do not screen the persons they watch yet.

Limits during the beta

Coverage of some PEP categories is still partial, in particular senior judges, senior military officers and the heads of state-owned companies.
PEP lists are the publishers’ own: some include local officials whose functions are not prominent public functions under the EU anti-money laundering rules.
Adverse media is not covered.
The screening section is not part of PDF extracts yet.

Getting Started

AI Agents

Integration Guides

Core Features

Behavior & Operations

Sanctions and PEP Screening

What is screened

How a match is read

When it arrives

Response

Fields

Where screening is not included

Limits during the beta

​What is screened

​How a match is read

​When it arrives

​Response

​Fields

​Where screening is not included

​Limits during the beta

What is screened

How a match is read

When it arrives

Response

Fields

Where screening is not included

Limits during the beta