Key finding
The most useful wire control is a record showing which number was dialed, who answered, and what was confirmed before release. Files that contain only an email approval cannot demonstrate independent verification.
Question and scope
This review looks at how settlement files document the verification of wiring instructions, and where the record becomes too thin to reconstruct the decision. The scope is the closing file: the instruction received, the callback attempted, the confirmation recorded, and the release approved. It does not evaluate any individual institution or estimate fraud rates.
The unit of observation is one disbursement event. A file can contain several wires, so each release is treated separately rather than counted once per transaction.
Evidence standard
A complete record contains the original instruction with its envelope details, the source of the phone number used, the call time, the identity of the person reached, the fields confirmed, and the name of the approver. A record missing the number source or the approver is incomplete even if the wire settled without incident.
Public guidance consistently recommends out-of-band verification of changes to payment instructions. The review uses that recommendation as a benchmark for the record, not as evidence that any particular file was fraudulent.
Observed failure patterns
Three patterns recur. First, a callback is made but the number comes from the same email that delivered the instructions. Second, the callback reaches a general line and the confirmation is recorded as verified without a named person. Third, a late instruction change is approved by reply to the originating thread.
Each pattern leaves the same gap: no independent party confirmed the account. The file may look complete because a note exists, but the note does not identify a separate source of truth.
Interpretation and limits
These observations describe documentation quality, not the prevalence of fraud. A file with a weak record may still involve a legitimate transfer, and a file with a strong record may involve a criminal who controlled both ends of the callback. The control reduces risk; it does not eliminate it.
The sample is also biased toward files that reached a review queue. Closing teams that only examine exceptions will see a different distribution than teams that sample completed wires.
Operational implications
Make the number source a required field in the wire record, and separate the preparer from the approver. Add a stop condition for any instruction that differs from a prior set, and require a new callback rather than an email clarification.
Review a monthly sample of completed wires for missing number sources and unnamed confirmations. Those two fields predict whether the file can survive a challenge better than the presence of a generic verification checkbox.