A SWIFT Compliance Checklist for Banks and Fintechs
A working checklist of the SWIFT-related compliance items that auditors, bank sponsors, and regulators actually ask about. Treat this as a starting point, not a substitute for jurisdiction-specific advice.
1. SWIFT Customer Security Programme (CSP)
- Annual self-attestation completed against the current CSCF (Customer Security Controls Framework).
- Independent assessment performed where required by your tier.
- Mandatory and advisory controls scoped, owned, and tracked in the security backlog.
- Architecture conforms to one of the supported types (A1/A2/A3/A4/B).
- SWIFT-related infrastructure segmented from general corporate network.
- Privileged access controls and MFA enforced on operator and admin accounts.
2. Sanctions screening
- Lists cover all jurisdictions touched by the bank/fintech (OFAC, UN, EU, UK HMT, local).
- List updates automated to the issuing authority's cadence.
- Matcher tuned with documented thresholds, reviewed quarterly.
- Both real-time screening (per-message) and batch re-screening (lists update against existing relationships).
- Review queue with documented SLA and audit trail.
3. AML/CFT monitoring
- Pattern-based transaction monitoring rules and/or models in production.
- Typologies updated against FATF and local FIU guidance.
- Alert disposition with documented investigative steps.
- Suspicious activity reporting workflow with required filing timelines.
4. ISO 20022 readiness
- In-scope cross-border flows live on MX (CBPR+ coexistence ended November 2025).
- Capture-side UX collects structured originator and beneficiary fields.
- Internal data model MX-shaped, not MT-shaped.
- Translation layer for any remaining legacy MT counterparties.
5. SWIFT gpi adoption
- gpi membership active for in-scope flows.
- UETR captured, persisted, and exposed to the relevant business users.
- gpi Tracker integrated into operational tooling.
- Customer-facing status surfaces use gpi data where available.
6. Audit trail and controls
- Append-only ledger of every SWIFT message sent, received, and acted upon.
- Maker-checker enforced on payment initiation above thresholds.
- Segregation of duties enforced by the platform, not policy.
- Annual control testing performed and documented.
- External audit walkthrough rehearsed at least annually.
7. Charge bearer and fee transparency
- Charge-bearer default (OUR/SHA/BEN) matched to use case.
- FX margin disclosed where regulator requires; transparent line-item by preference.
- gpi fee data surfaced where available.
8. Vendor and partner management
- Correspondent banks' compliance attestations on file.
- Service-provider security assessments current.
- Contractual right-to-audit clauses on critical vendors.
- Incident-notification clauses on payment-related vendors.
9. Incident response
- Documented runbook for SWIFT-related incidents (fraudulent message attempts, CSP control failures, gpi outages).
- Tested annually with a tabletop or full exercise.
- Notification paths to SWIFT, regulators, and sponsoring banks pre-defined.
10. Training
- Operator training on payment initiation, screening review, and incident response.
- Compliance training updated to current sanctions regimes and AML typologies.
- Engineering training on CSP architectural requirements.
Key takeaways
- Compliance is not a single document; it is the operating condition of the platform.
- The most expensive findings come from gaps the platform cannot demonstrate at click of a button.
- Build the controls in product, and the checklist becomes a status report rather than a project.
FAQ
Is this enough for a license application? No, license applications require jurisdiction-specific tailoring. This checklist is the operating layer underneath.
How often should this be reviewed? Quarterly, with the annual CSP attestation as a hard anchor.
Related reading
For emerging-market banks, SWIFT is not optional. The fragility is in the correspondents on either end of the message.
MT was a printer-line format. MX is structured data. The difference is the entire next decade of cross-border product.
Stablecoins solve a real cross-border problem in specific corridors. They do not solve every cross-border problem in every corridor.