CR360 Country Risk 360
CR360 Client Screening · Methodology v1.0

How CR360 Client Screening works

CR360 uses one simple client-screening journey with sector-appropriate risk models underneath it. The user supplies factual onboarding information; CR360 selects the applicable model from the product or service being provided, weights five risk domains, applies concentration and legal override rules, and produces an explainable point-in-time result.

One screening journey, four risk models

ML/TF/PF risk does not arise with identical significance across all business models. CR360 therefore does not use one universal weighting formula. Instead, the selected product or service automatically maps the assessment to one of four risk-model profiles. No additional institution-risk questionnaire is required during client screening.

The weights are CR360 methodology parameters, not regulator-prescribed percentages. They are intended as explainable defaults for SMEs and must remain consistent with the accountable institution's own entity-wide risk assessment, products, client base, channels and risk appetite.

CR360 risk-model profiles

DomainA · Relationship / ProtectionB · Advisory / StructuringC · Asset / Investment / High-ValueD · Transactional / Value Transfer
Customer / occupation25%30%20%20%
Source of funds20%25%30%15%
Product / service / relationship exposure25%20%25%25%
Geographic exposure15%15%15%25%
Delivery channel15%10%10%15%

Current product-to-profile mapping

ProfileTypical CR360 product/service selectionsMethodology rationale
A · Relationship / ProtectionInsurance; pension / retirementLonger-term or comparatively constrained relationships where customer and product characteristics matter but rapid value transfer is not the core purpose.
B · Advisory / StructuringProfessional / advisory service; Other (default pending contextual review)Relationships where client characteristics, purpose, structuring context and source of funds are comparatively important.
C · Asset / Investment / High-ValueInvestment / securities; property; high-value goods / assetsRelationships where origin of funds and the nature or value of the asset/service are primary risk drivers.
D · Transactional / Value TransferBanking/payment account; derivatives/CFDs; crypto; remittance/FXRelationships where portability, transaction velocity, remote access and cross-border exposure are comparatively more significant.

How the 100-point score is calculated

Each of the five domains is first classified LOW, MEDIUM or HIGH from factual client inputs. The applicable profile weight is then converted into a contribution:

Domain ratingContribution to that domain's weightExample where domain weight = 30%
LOW0%0 points
MEDIUM50%15 points
HIGH100%30 points

The five contributions are added to produce a 0–100 score. The applicable profile then determines the initial risk band:

ProfileLOWMEDIUMHIGH
A · Relationship / Protection< 3030 to < 6565+
B · Advisory / Structuring< 3030 to < 7070+
C · Asset / Investment / High-Value< 3030 to < 7070+
D · Transactional / Value Transfer< 4040 to < 7575+

Profile D uses a higher LOW/MEDIUM transition because normal transactional/value-transfer relationships can legitimately contain more medium-level characteristics (for example remote delivery or cross-border functionality) without being equivalent to a medium-risk relationship in a comparatively constrained protection model.

How geographic risk is treated

The numerical geographic domain uses the client's country of residence / domicile and the current CR360 jurisdiction classification. Country of birth remains recorded for identity resolution, sanctions/PEP screening context and audit evidence, but it no longer contributes numerical points merely because the person was born there.

CR360 jurisdiction ratingClient-domain ratingTreatment
LOWLOW0% of the applicable geographic-domain weight
MEDIUMMEDIUM50% of the applicable geographic-domain weight
HIGHHIGH100% of the applicable geographic-domain weight
EXTREMEHIGH + mandatory escalation100% of the applicable geographic-domain weight plus separate escalation

Country classifications are derived from the current CR360 jurisdiction methodology.

Customer / occupation risk

Employment status and industry/sector are converted to a LOW, MEDIUM or HIGH customer-domain rating. The user supplies facts rather than choosing the risk category. Employment status may increase the sector result where the economic profile is unclear or presents additional risk.

Source of funds

Source of funds means the origin of the money expected to be used in the relationship or transaction. Salary, ordinary self-employment/business income and pension-type sources are normally treated LOW; savings, investments and documented asset/inheritance sources are normally MEDIUM; third-party funding, unsupported inheritance, opaque private transfers and other complex sources are normally HIGH. Documentary verification remains the institution's responsibility.

Source of wealth is captured as contextual information and becomes particularly important where EDD is required.

Product / service risk

The product selection first selects the relevant risk-model profile. The Product / service domain then combines the product baseline with the expected monthly activity and expected maximum transaction already captured in the screening form. This allows the domain to move from LOW to MEDIUM or HIGH without adding another client question, while avoiding double-counting an entire sector as automatically high risk.

Delivery-channel risk

Face-to-face onboarding with independent verification is treated LOW; remote onboarding with strong independent identity controls is MEDIUM; remote or indirect onboarding with materially weaker independent verification is HIGH.

Concentration, boundary and override controls

Arithmetic cannot be allowed to dilute a concentrated risk pattern. CR360 therefore applies the following controls after the initial numerical band is calculated:

  • Any HIGH domain prevents an overall LOW outcome.
  • Three or more HIGH domains force an overall HIGH outcome.
  • Two or more HIGH domains force HIGH where those domains represent at least 50% of the applicable profile weight.
  • A score within 5 points of either profile threshold is flagged as a borderline score and reviewer confirmation is recommended.
  • A declared foreign PEP creates a HIGH-risk override and EDD treatment.
  • An EXTREME country-of-residence exposure creates a HIGH-risk override and mandatory escalation.
  • Sanctions/TFS matches remain legal control events and are not diluted into the numerical score.

Model calibration and stress testing

CR360 tested the v1.0 model before implementation rather than selecting thresholds solely for mathematical convenience.

400,000synthetic client assessments across 100 Monte Carlo calibration runs
972exhaustive domain states (all 3⁵ combinations across four profiles)
0monotonicity violations after the final concentration rule
200,000one-domain error simulations used for resilience testing

Portfolio calibration result

ProfileLOWMEDIUMHIGH
A · Relationship / Protection68.7%22.0%9.2%
B · Advisory / Structuring68.9%21.7%9.4%
C · Asset / Investment / High-Value63.6%26.7%9.7%
D · Transactional / Value Transfer61.2%29.1%9.7%

These are synthetic calibration results, not forecasts of the risk distribution of any real customer book. The implementation-specific test population used an underlying 70% LOW / 20% MEDIUM / 10% HIGH risk mix as a portfolio sanity benchmark and then generated the same product/activity inputs used by the live engine. CR360 was not instructed to reproduce that final distribution. Profiles C and D naturally produced more MEDIUM outcomes because ordinary asset/value-transfer relationships can carry more medium-level product/activity characteristics.

Adversarial robustness

StressABCD
Mean band change after random ±20% weight perturbation3.27%2.62%2.68%1.98%
Mean band change after ±5-point threshold stress4.39%2.88%2.88%1.65%
One-domain error causing MEDIUM/HIGH → LOW downgrade2.77%2.65%2.71%4.06%

The adversarial state-space test initially identified a weakness where three HIGH domains could remain MEDIUM in certain B/C combinations because the selected domains represented less than 50% of the model weight. The rule was strengthened so that three or more HIGH domains always force HIGH. The complete state space was then rerun with zero such failures and zero monotonicity violations.

Validation is ongoing. CR360 will rerun calibration and stress testing when material weights, thresholds, domain mappings or override rules change. Historical reports record the methodology version applied.

How does sanctions screening work?

CR360 screens natural-person names against authoritative targeted-financial-sanctions data made available by the South African Financial Intelligence Centre and the United Nations Security Council. Name matching is conservative and potential matches must be resolved by a human using the underlying identifiers before onboarding proceeds.

CR360 does not automatically confirm a sanctions designation. Potential matches are surfaced for human resolution; a confirmed applicable designation is a legal control event rather than a score contribution.

PEP / prominent-person treatment

CR360 combines the client's declaration with targeted open-source screening. A declared foreign PEP creates a HIGH-risk override and EDD recommendation. Domestic prominent-person status is sent for risk-based review rather than automatically treated as HIGH. Family members and known close associates require review of the linked relationship.

Adverse-information screening

CR360 runs controlled name-and-context searches for potentially material financial-crime, legal/regulatory and public-office indicators. Identity confidence and risk relevance are assessed separately. Surfaced results are labelled for review rather than treated as findings of wrongdoing.

Risk-based onboarding controls

The purpose of the rating is operational: higher-risk relationships receive stronger, more frequent and more resource-intensive controls while lower-risk relationships remain subject to baseline CDD. CR360 uses the following recommended review defaults:

RiskFile review defaultProof-of-address freshness default
LOWEvery 3 yearsWithin 12 months where POA forms part of the institution's verification method
MEDIUMEvery 2 yearsWithin 6 months
HIGHAt least annuallyWithin 3 months
These are CR360 methodology defaults, not universal statutory time limits. The accountable institution's RMCP remains controlling.

Methodology basis and sources

The model is designed to support a documented risk-based approach in which client/customer, product/service, delivery-channel and geographic indicators are assessed together and the institution can explain how the indicators are weighted and combined. CR360's public methodology should be read together with the institution's own RMCP and entity-wide risk assessment.

Regulatory reference points used in the design include FIC Guidance Note 7B and PCC 53. Guidance Note 7B requires accountable institutions to document their risk-rating methodology and procedures in the RMCP. Official sanctions sources include the FIC Targeted Financial Sanctions facility and the UN Security Council Consolidated List.

Does CR360 approve or reject the client?

No. CR360 provides decision-support. The accountable institution remains responsible for its RMCP, customer due diligence, risk appetite, sanctions-match resolution, enhanced due diligence, transaction monitoring and final onboarding or continuation decision.

Back to top ↑