Risk-Based Client Due Diligence Matrix
Risk-Based Client Due Diligence Matrix
AML/CFT Onboarding Risk Scoring – Project Summary
Prepared by: Tommy O’Brien
Role: Compliance Officer
Date: July 21, 2026
Page 1
Risk-Based Client Due Diligence Matrix
1. Risk Matrix Build
The workbook AML_Risk_Matrix.xlsx contains two linked sheets. Scoring holds the policy inputs (point
values and tier thresholds) that Compliance controls. Clients holds the client-level data and pulls every
point value and tier from Scoring using formulas, so changing a weight in one place recalculates the
whole matrix automatically.
Scoring Sheet
Figure 1. Scoring sheet — point mapping, volume bands, tier thresholds, and the override rule.
Clients Sheet (Computed)
Figure 2. Clients sheet — CountryPts, VolumePts, ProductPts, ChannelPts, TotalScore, Tier, and Rationale all
computed by formula.
Key formulas: CountryPts/ProductPts/ChannelPts use INDEX/MATCH against the Scoring point table;
VolumePts applies the <10k / 10k–50k / >50k band logic; TotalScore sums the four point columns; Due
Diligence Tier applies the 4–7 / 8–13 / 14–20 thresholds and is overridden to EDD whenever Country
Risk = High and Monthly Volume exceeds 50,000 USD, even if the raw score would fall lower.
2. Results Summary
Each client's automatically computed score, tier, and one-line rationale, ready to paste into a client file:
Page 2
Risk-Based Client Due Diligence Matrix
Client TotalScore Tier Rationale
Low-risk on every factor (country, volume, product,
AsterPay GmbH 4 SDD
channel), well under the SDD threshold.
Baltic Export Sp. z Medium country and product risk plus a mid-range
10 Standard
o.o. (10k–50k) volume land it squarely in Standard.
High country risk combined with medium volume,
Cedar Trading LLC 14 EDD product, and channel risk pushes the score into EDD
(score-based, not override).
Maximum risk on all four factors; also trips the High-
Dune Remit OU 20 EDD country + >50k-volume override rule independently of
the score.
High volume (>50k) is offset by Low country risk, so it
Echo Market SAS 12 Standard
stays in Standard rather than triggering the override.
Population split: 1 of 5 clients (20%) fall into SDD, 2 of 5 (40%) into Standard, and 2 of 5 (40%) into
EDD.
3. Calibration Reflection
● Is the EDD population reasonable? With a small sample of five clients, 40% landing in EDD
(Cedar Trading LLC and Dune Remit OU) is high for a genuinely risk-based approach — a well-
calibrated model should generally flag a minority of the book, not two-fifths of it. On a larger,
more representative client base this proportion would likely be lower, but it is still worth checking
that the weights aren't pulling too many mid-risk clients over the line.
● One weight/threshold I'd tweak: Cedar Trading LLC reaches EDD purely on TotalScore (14)
despite only Medium volume/product/channel risk — it is High country risk alone (5 points)
combined with three Medium ratings (3+3+3=9) that tips it over. The single largest lever is
Country Risk: dropping High-country points from 5 to 4 would move Cedar's score to 13
(Standard) without touching any other client's tier, since it is the only client sitting right at the
13/14 boundary.
● Which client(s) are EDD, and why: Cedar Trading LLC (score 14, no override triggered —
High country risk plus Medium ratings elsewhere) and Dune Remit OU (score 20, and
separately triggers the mandatory override because it is High-country with monthly volume over
50,000 USD). Dune Remit OU is the clearer EDD case: it is High risk on every single factor, so
both the score-based rule and the override agree.
● Value of the override rule: The override rule is doing useful, independent work: Dune Remit
OU would have hit EDD on score alone (20), so the override is redundant there, but it exists as
a backstop for a future client that is High-country and high-volume yet happens to have Low
product/channel risk pulling the raw score under 14 — exactly the kind of case a purely additive
score could under-flag.
● Next step: Before rolling this out, I'd re-run the matrix against a larger historical sample (50–100
onboarded clients) to confirm the EDD rate settles closer to an expected 10–15% of the book,
and I'd document the Country Risk weight change (if adopted) as a dated policy update in the
Scoring sheet's notes so the change is auditable.
Page 3