The single rulebook harmonises what you must do about money laundering. It does nothing to harmonise the lists you screen against — and that gap is a data problem, not a legal one.
The EU’s Anti-Money Laundering Regulation, Regulation (EU) 2024/1624, applies from 10 July 2027 and pulls customer due diligence, beneficial ownership and internal-controls obligations into one directly applicable text across the Union. The new Authority for Anti-Money Laundering and Countering the Financing of Terrorism, based in Frankfurt and operational since 1 July 2025, will supervise the highest-risk firms directly. A great deal will be genuinely harmonised. Your watchlists will not be. Sanctions and politically-exposed-person screening still draws on a scatter of separately maintained sources, each with its own scope, format, update cadence and legal authority, and reconciling them into one coherent screening decision is left entirely to you.
The rulebook harmonises obligations, not lists
It is worth being precise about what changes and what does not. AML/CFT and sanctions are legally distinct regimes. Restrictive measures flow from the EU’s Common Foreign and Security Policy and from national and third-country sanctions authorities; anti-money-laundering obligations flow from the AML package. The AMLR expects obliged entities to have policies, controls and procedures that ensure targeted financial sanctions are actually applied, but it does not publish the names. The names come from elsewhere, and “elsewhere” is plural.
Free · 4 minutes
When two of your systems disagree, do you know which one to believe?
Fourteen questions on ownership, lineage, and quality — the difference between a number on a dashboard and a number you could defend. Banded finding on screen, full sheet by email.
So the harmonisation you get from 2027 is at the obligation layer — the same due-diligence rules, the same beneficial-ownership thresholds, the same expectation that screening works. The fragmentation you keep is at the data layer. A firm that reads the AMLR as “the lists are now sorted for us” has misread it. If anything, a single supervisor asking consistent questions across the Union raises the bar on being able to show that your list handling is coherent. That is the same shift I described in what AMLA’s direct supervision and its data requests will look like: the obligation is harmonised, the evidence is on you.
Four list families, four data shapes
Most EU firms screen against at least four families of source, and each behaves differently:
- The EU consolidated list. The Commission maintains a consolidated list of persons, groups and entities subject to CFSP-related financial sanctions — the Financial Sanctions Files — distributed in a structured XML format for exactly this purpose. The important caveat: only the measures as published in the Official Journal are legally authentic. The convenience file is a mirror, not the source of truth.
- The UN Security Council Consolidated List. Distributed as its own XML, on its own schedule, with its own identifier conventions. Much of it is reflected into EU measures, but not instantly and not always identically.
- Third-country lists with a nexus to your business. The US OFAC Specially Designated Nationals list and the UK OFSI consolidated list matter wherever you touch USD clearing, correspondents or UK-nexus clients. These are not EU obligations, but ignoring them is a commercial and legal risk of a different kind.
- PEP data. Not a sanctions list at all. The EU publishes a list of the prominent public functions that make a person a PEP, not a register of the people. The actual person-level data comes from commercial providers, and it is probabilistic, dated and duplicative in a way sanctions lists are not.
Treat those four as one undifferentiated blob and you get the two failure modes that a screening programme exists to avoid: the same underlying person raising three near-identical alerts because they sit on three lists, and a true match slipping through because one source lagged, changed its identifier scheme, or was ingested and silently dropped.
Design the reconciliation, do not improvise it
The screening data layer has one job before any name is ever matched: take heterogeneous sources and produce a single normalised, versioned, provenance-tagged set of screening targets. Three properties make it defensible.
Provenance survives normalisation. Every target record carries which list it came from, the list’s published version or date, and the original per-list identifier. When an analyst clears an alert, they need to see that this is the EU designation and the UN designation of the same individual, not two separate coincidences. When a designation is lifted, you need to remove it from the right source without deleting a still-live match from another.
Overlaps are resolved, not hidden. A person on the EU, UN and OFAC lists should resolve to one screening entity with three linked designations. That is what collapses three alerts into one review while preserving the fact that three regimes bite. Getting this right is an entity-resolution problem, and it is the part firms most often skip.
The pipeline is auditable end to end. You must be able to state which version of which list was live at the moment a given transaction or onboarding was screened. This is the data-lineage requirement the whole exercise turns on, and it is the same discipline I argued for in re-architecting sanctions screening for instant payments: screen against a known, dated snapshot, not against “whatever the feed happened to hold”.
A normalisation pass, concretely
Below is the shape of the ingestion-and-normalisation step: read each source in its native form, map it onto one canonical schema, and key overlapping designations to a single resolved entity. It is deliberately source-agnostic — the per-list adapters vary, the canonical record does not.
from dataclasses import dataclass, field
from datetime import date
@dataclass(frozen=True)
class Designation:
source: str # 'EU_FSF', 'UN_SC', 'OFAC_SDN', 'OFSI'
source_id: str # the list's own identifier
list_version: str # published version or date of the snapshot
names: tuple # primary name + aliases, normalised
programmes: tuple # e.g. ('RUSSIA',) — the regime that bites
@dataclass
class ScreeningEntity:
entity_key: str # resolved cross-list key
designations: list = field(default_factory=list)
def ingest(adapter, raw, snapshot: str) -> list:
# Each adapter parses one source (XML, CSV) into Designations.
# snapshot pins the exact list version screened against.
return [adapter.to_designation(row, snapshot) for row in raw]
def reconcile(designations: list, resolve_key) -> dict:
# resolve_key groups EU/UN/OFAC records for the same target
# into one entity, preserving each designation and its provenance.
entities: dict = {}
for d in designations:
key = resolve_key(d) # deterministic + fuzzy tier
entity = entities.setdefault(key, ScreeningEntity(key))
entity.designations.append(d) # keep all sources, do not merge away
return entities
# One person on three lists -> one entity, three linked designations.
# One alert to review; three regimes still evidenced.
The point is not the twenty lines. It is that reconciliation is an explicit, tested stage with its own outputs, not a side-effect of whatever your screening vendor does internally where you cannot see it.
Matching that is neither noisy nor blind
Once targets are normalised, matching thresholds should be tuned per source and per data quality, not set once globally. Sanctions designations, being authoritative and comparatively clean, justify tighter, more aggressive matching — a missed true match here is the worst outcome. Commercial PEP data, being fuzzier, needs looser matching paired with risk-based handling so it does not drown analysts. A single global threshold optimised for one of those is wrong for the other, and that single decision is often what separates a screening function that surfaces real risk from one that buries it under noise. This is the configurability point I made about turning the single rulebook into a controls engine: the obligation is fixed, the tuning is yours to justify.
What a supervisor will ask to see
When AMLA or your national supervisor tests this, the questions are concrete. Which lists do you screen, from which authoritative source, at what refresh interval? Show the snapshot that was live when this customer was onboarded. Show how the same designated person on three lists produces one review, not three. Show how a delisting propagated. Show why your PEP threshold differs from your sanctions threshold and who signed that off. None of that is answerable from a screening tool’s dashboard alone; it is answerable from a data layer you designed to be asked these questions.
The single rulebook will tempt firms into thinking the hard part is now centralised. The hard part was never the obligation. It was always the lists — and those are still yours to reconcile, one snapshot at a time.
Free interactive tool
Website compliance checklist
What your site has to do, based on what it actually does
Answer as much or as little as you like — the list builds as you go. Nothing is stored against your name and no email is required.
Everything that applies
Ordered by what to do first: legal requirements you can close quickly, then larger pieces of work, then what is expected rather than required. Not exhaustive, and not a legal audit.
Dated PDF, yours to keep or circulate.
Free interactive tool
Interactive deadline calculator
Check which regulations apply to you and when
Regulation across the EU, UK, US and Asia-Pacific has moved considerably in the past eighteen months, and several headline dates have shifted more than once. Twelve questions, about three minutes.
Results are shown on screen — no email required. A dated summary is available to download, and can be sent on if that's more useful. What we do with your answers.
Most technology problems are not technology problems. They are control problems.
The systems exist. The investment has been made. The question is whether leadership can understand, direct, evidence, and sustain what those systems produce. Find out where control exists — and where it only appears to.
Full Governance by Sixteen Pillars
Govern your business. Prove your compliance.
A board assurance cockpit for EU-regulated financial firms — tamper-evident, hash-chained proof of governance across DORA, GDPR, NIS2, ISO 27001, the EU AI Act and MiCA. In development.
See what's coming