Every regulated firm must screen customers against sanctions lists, politically exposed person records and adverse media. Almost none builds that capability itself, because the work is not the matching logic but the data: who compiles it, how fast it updates, and how reliably a name in your file resolves to a person in theirs.

Four things to consider

Coverage against your actual exposure. Headline list counts matter less than whether a provider covers the regimes and jurisdictions your investors touch, including local lists beyond OFAC, the EU, HM Treasury and the UN.

Refresh frequency. A sanctions designation takes effect on publication. Ask how long a new designation takes to reach your screening, and whether that figure is measured or aspirational.

Entity resolution quality. The cost of screening is analyst hours spent clearing alerts, not the licence. Ask how the provider handles transliteration, name order and common names.

Adverse media scope. Adverse media has no authoritative list, so sourcing defines what you see. Ask which sources are included, how allegations are distinguished from convictions, and how duplicates are handled.

The best PEP and sanctions screening vendors

1. World-Check

A subscription database of PEP, sanctions and adverse media profiles delivered by API, and one of the longest established names in the category. Part of LSEG.

• Around 5.8 million person and organisation profiles

• REST and JSON API with real-time and delta update options

• Research analysts covering more than 70 languages

2. ComplyAdvantage

The clearest example of a vendor that is both things: it compiles its own risk intelligence and sells a screening and monitoring platform on top of it.

• Sanctions coverage extending beyond OFAC, HM Treasury and the EU to
  more than 60 jurisdictions

• Data it describes as refreshed in minutes rather than months

• Coverage across 14 languages

• Agentic AI it claims resolves up to 85% of routine alerts

3. Dow Jones Risk & Compliance

Licensable compliance data content, built for integration into third party screening platforms rather than sold as an application.

• PEP records dating to 2001 and adverse media entities from 2012

• Several hundred researchers working across roughly 60 languages

• Backed by the Factiva archive, adding more than 600,000 articles daily
  from 33,000 sources

• Internally measured average precision of 99.74% for the twelve months
  to July 2024

4. Sumsub

An identity verification and onboarding platform whose screening sits
alongside document verification, KYB and ongoing monitoring.

• Screening across a claimed 50,000 data sources in more than 240
  countries and territories

• List coverage including OFAC, the UN Security Council, HM Treasury, the
  EU consolidated list and Australia's DFAT

• Adverse media monitoring with assisted match resolution

• Screening data licensed from ComplyAdvantage and World-Check One

How to run a screening proof of concept

Coverage claims cannot be compared from a brochure, because no two vendors count the same things the same way. Test with your own data.

Take a sample of your live book, including the awkward cases: transliterated names, common names, entities in jurisdictions with thin public records, anyone previously flagged. Run the same sample through each candidate at the same settings.

Then compare four things:

• Hit counts, which tell you how much work each vendor creates

• True positives, checked against what you already know about your
  book

• False positives, which are the real running cost

• Adjudication effort, the time an analyst needs to clear a typical
  alert

More hits is not better screening, and fewer is not safer. What matters is the ratio and what it costs to work through it, which is also why reducing false positives in AML screening deserves its own proof of concept rather than being treated as a side effect of picking a vendor.

Screening inside the onboarding workflow, with Steward

Screening rarely fails because a name was missed. It fails because alerts arrive faster than a team can clear them, and the reasoning behind each decision is not recorded well enough to stand up later, a pattern that runs through why AML compliance breaks down operationally well beyond screening alone.

Steward is an AI-first AML and KYC platform for investor onboarding. Screening against sanctions, PEP and adverse media data sits inside the onboarding workflow, whatever provider you decide to pick, rather than beside it, so a result attaches to the investor file it belongs to.

If clearing alerts is where your onboarding time goes, book a demo.