AI OFAC Sanctions Screening & Counterparty Risk Assessment Playbook

Your compliance team has received an alert on a $4.8M wire transfer from a correspondent bank in the UAE. The originator name partially matches an SDN list entry, but the account number and country codes differ. You have 24 hours to clear, block, or reject before the Fed's deadline. Legal and compliance need a defensible written analysis.

When to use this playbook

• Use this playbook when the decision looks like the situation above: Your compliance team has received an alert on a $4.8M wire transfer from a correspondent bank in the UAE.
• It is a fit when you have source files in hand and need a structured, reviewable analysis — not a generic chat answer about "OFAC Sanctions Screening & Counterparty Risk Assessment".
• Do not use it as a substitute for licensed, legal, clinical, or authorized official judgment in the domain.

What you'll need

• Source documents specified in the workflow

You are an OFAC compliance analyst conducting an expedited sanctions screening review on a flagged international wire transfer.  I am attaching: - Wire transfer SWIFT message (MT103) - OFAC SDN and Consolidated Sanctions List extract for the flagged name - Correspondent bank CDD file for the originating institution - Prior transaction history with this counterparty

Work only from the attached source files. If a conclusion is not supported, say so.

Produce:
1. Perform a structured name-matching analysis comparing the originator name in the MT103 against all relevant SDN entries — accounting for transliteration variants, name order differences, and common aliases.
2. Assess whether the differences in account number and country code constitute sufficient basis to clear the match, or whether residual risk requires escalation.
3. Review the correspondent bank CDD file for any prior sanctions-related findings, geographic risk factors, or AML deficiencies noted by regulators.
4. Analyze the prior transaction history with this counterparty for pattern consistency — flag any changes in transaction size, frequency, or routing that are inconsistent with the established relationship.
5. Produce a written OFAC determination memo documenting the analysis, the decision rationale, and the evidentiary basis — suitable for examiner review.

Call out where independent models are likely to disagree, and list follow-up documents a reviewer should request.

What to expect

• Multi-model consensus name-match scoring with confidence percentage
• Residual risk assessment with clear/block/escalate recommendation
• Correspondent bank risk flag register
• Transaction pattern anomaly analysis
• Draft OFAC determination memo with model-agreement score

Review before you act

• Validate this output against source files before relying on it: Perform a structured name-matching analysis comparing the originator name in the MT103 against all relevant SDN entries — accounting for transliteration variants, name order differences, and common aliases.
• Validate this output against source files before relying on it: Assess whether the differences in account number and country code constitute sufficient basis to clear the match, or whether residual risk requires escalation.
• Validate this output against source files before relying on it: Review the correspondent bank CDD file for any prior sanctions-related findings, geographic risk factors, or AML deficiencies noted by regulators.
• Validate this output against source files before relying on it: Analyze the prior transaction history with this counterparty for pattern consistency — flag any changes in transaction size, frequency, or routing that are inconsistent with the established relationship.
• Confirm every cited figure, date, counterparty, or requirement against the attached originals — models compress and can drop a qualifier.
• Treat disagreement between models as a review item, especially on classification, materiality, and recommended next action.
• Do not authorize an operational, clinical, legal, credit, or enforcement action solely because the models agree.