Payment Routing: Preferred Does Not Mean Mandatory
Lumina Academy
When SwiftRef and a customer name different intermediaries, validate the route before overriding it; preferred does not mean mandatory.
Payment BA Awareness — Episode 44
SwiftRef says Chase New York, but the customer’s instruction says Citi New York. Should the payment engine simply replace Citi with the preferred intermediary? My first question is not which field to populate. It is whether “preferred” has been confused with “mandatory”, and what the customer’s instruction means for this payment.
Why the difference matters
Consider a USD payment to Beneficiary Bank B. The SSI or SwiftRef data points to Chase New York, while the customer has supplied Citi New York. Both names are present in the discussion, but their sources serve different purposes. I cannot assume that reference data should always win just because it is available to the engine.
The beneficiary may have pre-advised a particular USD Nostro account. The beneficiary account may also belong to an SPV or dedicated structure that needs a specific route. These are possibilities to investigate, not facts I would assume for every payment. They explain why replacing a customer-supplied intermediary without validation can create delays, repairs or unreconciled credits.
Before writing the routing requirement, I would make the conflict explicit:
• Customer instruction: Citi New York.
• SSI or SwiftRef reference: Chase New York.
• Business question: is one route preferred, or is a particular route mandatory?
• Decision required: preserve, derive, repair or seek confirmation.
A field-population rule is not complete until we know which of those decisions applies and why. Otherwise, the system may perform exactly as specified while failing to respect the intended payment route.
Scenario one: no intermediary is supplied
When the customer has not supplied an intermediary, reference data can fill a genuine gap. The bank can use the beneficiary bank, payment currency, internal routing and SSI or SwiftRef information to derive the route. Here, the requirement is about obtaining information that is missing.
I would still want the BA discussion to explain which source is used and how it relates to the beneficiary and currency. “Derive intermediary” is too short if different sources can produce different results. The requirement needs enough clarity for someone to understand the selection rather than simply observe a populated field.
The important point is that there is no customer-supplied intermediary to overwrite in this scenario. That makes it a different decision from the next case, even if the same reference-data service is involved.
Scenario two: the customer supplies an intermediary
If the customer supplies Citi New York, I would use reference data to validate the route rather than automatically replace it. A difference from the preferred route is a reason to ask questions, not enough evidence on its own to discard the instruction.
My validation questions would be:
1. Is this intermediary valid for the beneficiary bank and currency?
2. Is it on the customer’s agreed routing list?
3. Can our bank settle through the proposed route?
4. Does the instruction conflict with mandatory routing requirements?
If the supplied route is valid, preserve the customer’s route. If it is unclear or invalid, move to repair or obtain customer confirmation before processing. I would not conceal uncertainty by silently substituting the reference-data preference.
This is where the requirement needs precedence rules. Customer instructions, agreed customer routing, internal routing and SSI or other reference data may disagree. The BA must establish how those sources relate to one another; I am not proposing a universal hierarchy that applies to every bank and payment.
So, What Makes the Difference?
For me, a good BA separates a missing value from a conflicting value. That leads to more useful questions and a clearer decision:
• Ask whether the rule is preferred or mandatory before implementing an override.
• Distinguish deriving a route from validating a supplied route.
• Consider the beneficiary’s funding and account context without inventing assumptions.
• Establish precedence when valid routing sources disagree.
• Make repair or customer confirmation explicit when the route is uncertain.
I would also challenge the acceptance criteria with both scenarios. A solution that derives correctly when nothing is supplied may still behave incorrectly when the customer has supplied a valid alternative. Testing only the empty-field case would miss the central issue in this episode.
Key Takeaways
• PREFERRED IS NOT MANDATORY: A reference preference is not automatically a compulsory route.
• FILL GAPS, VALIDATE INSTRUCTIONS: Missing and supplied intermediaries need different decisions.
• DO NOT SILENTLY OVERRIDE: Preserve a valid customer route; investigate unclear or invalid instructions.
• DEFINE PRECEDENCE: Requirements must address disagreement between routing sources.
At Lumina Academy, I use questions like these to move beyond message fields into practical payments BA thinking. For a one-to-one discussion, visit https://topmate.io/jebaraj_d.
Changed: Restructured Episode 44 into a blog, retaining the Chase New York versus Citi New York example and the distinction between derivation and validation.
More stories
- Late consent after a payment recall
- Why payments need multiple statuses
- When the expected Nostro debit is missing
- Which STP failures belong in repair?
- Preferred versus mandatory intermediaries
- Should a failed cover payment be recredited?
- How banks set a payment cut-off
- When does camt.055 trigger camt.056?
- INDA vs INGA — One Code, Two Accounting Outcomes 🤔
- pacs.028: A Missing pacs.002 Does Not Mean Payment Failure
- pacs.004 FOCR: Check Both Cover Legs Before a Refund
- pacs.009 COV Rejected with FF01: Do Not Rush the Refund
- Payment Cut-Off Times: Work Backwards from the Deadline
- camt.055 to camt.056: Check the Payment Lifecycle First
