Due Diligence 12 November 2026 · Gumshoe

Xero Flags the Bank Change. Who Actually Checks It?

An email arrives from a supplier you've paid a dozen times. The logo is right. The invoice number follows the sequence. Only the BSB and account number are new.

An email arrives from a supplier you've paid a dozen times. The logo is right. The invoice number follows the sequence. Only the BSB and account number are new.

The technology flagged the change. The process let it through.

The gap isn't technical. Xero has tracked a product idea for routing supplier bank-change notifications to specific users. Users on that thread point out that staff bank changes alert every payroll admin automatically. Supplier changes? Only the person who made the edit gets notified. That asymmetry is the control failure.

Your accounting software will tell you a bank account changed. It will not verify the change was legitimate. That step — the callback to a known number, the second sign-off, the documented evidence — lives entirely outside the platform. Until it's built into your process, every alert is just noise.

The Illusion of Notification: How Bank Changes Are Currently Flagged

Xero's notification logic treats a supplier's bank account change as a private note to the editor — not a risk event requiring oversight. When a staff member updates their own bank details, every payroll admin receives an automatic alert. When a supplier's bank details change, only the user who made the edit is notified. No secondary reviewer. No approval workflow. No audit trail that anyone else can see without digging through the history log.

This asymmetry is the default, not a bug. The platform assumes the person entering the change is the control. In practice, that person is often the only one who sees it.

The control gap isn't theoretical. Users on the product forum consistently note the asymmetry: payroll admins are looped in automatically when an employee updates their BSB, but a supplier's bank change — often for far larger sums — triggers a single email to the person who typed it. That person might be a junior bookkeeper, a temp or, if a login has been compromised, the fraudster. The system treats the edit as a data entry task, not a treasury event.

Until Xero changes this, the workaround is manual: someone other than the editor reviews every supplier bank change and documents the callback. Every step lives outside the software.

The Verification Chasm: From Digital Flag to Hard Stop

A notification is not a verification. The email lands. Someone reads it. The invoice gets paid. The control exists only in the audit trail, not in the moment the money moves. Xero's contact history shows who edited the record and when. It does not show who called the supplier, confirmed the BSB against a known statement, or signed off before the next payment run.

The gap is structural. The platform treats a supplier bank update as a record-keeping event: update the contact, save, done. It offers no native "four-eyes" gate, no mandatory callback field, no workflow that blocks payment until a second authorised user confirms the change was verified independently. The alert is passive. The verification — if it happens — lives entirely in your procedures, not in the software.

Most teams mistake the alert for the control. They build a process around "check the email" when the real control is "call the supplier on a known number, match the BSB to a statement, document the callback, then approve." The software cannot enforce that sequence. It can only record that the record changed. The hard stop — the point where a payment run refuses to execute because the verification field is blank — is not something Xero offers: there is no built-in step that holds a payment until a bank change has been checked. It exists only in your operating procedure.

That handoff, between the notification and the next payment batch, is where payment redirection losses happen. The control is the callback you made, the statement you matched, the sign-off you documented. Everything else is just a log entry.

Deconstructing Controls: Segregation Beyond the Workflow

The person who types the new BSB cannot be the person who clicks "pay". That is not a software setting — it is a structural rule. When the same login updates a supplier's bank details and authorises the next payment run, the control environment has collapsed into a single point of failure. No audit trail, no approval matrix, no notification routing can compensate for that concentration of authority.

Under the Corporations Act 2001, directors and officers must exercise reasonable care and diligence. A process that lets one person change where a supplier gets paid, with no independent check, is hard to defend against that standard. The software does not enforce segregation. It records the change and emails the person who made it. The segregation lives in your roster, your approval matrix, and the discipline that stops the updater from also being the authoriser.

The asymmetry matters here too. Staff bank changes notify every payroll admin; supplier changes notify only the updater. That design assumes the updater can be trusted. A fraud control can't make that assumption.

Close the loop with a hard rule: the credential that writes the bank record must be different from the credential that releases the funds. Enforce it through role-based permissions, dual-authorisation payment batches, or a mandatory second-signature checklist — whatever your system supports. If your accounting package cannot enforce that separation, the control sits in your procedures manual and the disciplinary consequence for bypassing it. Anything less is not a control; it is a hope.

Practical Guardrails: Controls for Bookkeepers and Practices

Start with what you can set today. Limit who can edit supplier contact details to as few users as practical, and keep that list separate from the people who approve payments. Then turn on dual authorisation for payments in your bank's business banking portal, so every batch needs two approvers and neither is the person who keyed the invoices. Write the rule into your engagement letter or procedures manual.

Add a weekly review: a partner or director goes through every supplier bank change made that week (each contact's history shows who changed what and when) and signs off each one before the next pay run. It uses what you already have, costs nothing extra, and leaves a paper trail an auditor can follow. But it still relies on people, after the fact.

Verification tools take the check out of the inbox. Gumshoe, our own service, checks the supplier's ABN and company records against Australian government registers and finds the business's phone number from independent sources, so the callback goes to a number the fraudster didn't supply. On paid plans it can also place the verification call and record the result. Whatever you use, the cost is small next to a single redirected payment.

If you manage multiple clients, standardise the control across every file. Build the verification step into your month-end checklist template: "Supplier bank changes validated — Y/N, by whom, tool used." A partner signs off the checklist, not the individual bookkeeper. That single column turns a scattered habit into a firm-wide control a reviewer can trace quickly.

Where most practices stall is between manual review and full automation. Set a hard rule: every supplier bank change triggers a callback to a known phone number, not the one on the invoice or in the email, before the next payment run. Log the call, the person you spoke to and the outcome in the contact notes, so the record sits next to the change. It costs nothing.

Uncommon Insights

Notification routing has been an open request for more than a decade. That isn't only a product gap. The risk sits outside the ledger: in the decision to trust a changed record without an independent check.

Most controls budgets favour detection: reconciliations, exception reports, post-payment audits. With payment redirection, detection often comes too late, because money that has reached a fraudster's account can be hard to recover. Preventive controls, a callback and a second approver on master-file changes, cost far less than a single fraudulent payment.

The callback is the cheapest control in the stack. A short call to a known number, not the one on the invoice and not the one in the email thread, breaks the attack before the payment file is generated. Yet many practices treat it as optional, something to do when something feels off. That discretion is the weakness. Make the call for every bank account change, every time, and log it where the change is recorded. Software vendors optimise for workflow efficiency; fraudsters exploit the gap between efficiency and verification. Close it with procedure.

Key Takeaways

  • Mandate callback verification for every supplier bank account change — a short call to a known, independently sourced number breaks the fraud chain before the payment file is generated.
  • Enforce segregation of duties on master-file edits — the person who updates bank details must never be the person who authorises the payment; use your bank's dual-authorisation settings and tight user permissions.
  • Have someone other than the editor review every supplier bank change — Xero only notifies the person who made it, so build the second pair of eyes into your weekly routine.
  • Log every verification in the same system that records the change — an auditable trail linking the alert, the callback, and the approver gives you evidence that the check was done.
VERIFY NOW

Run a free supplier check in seconds

Search by business name, ABN, or ACN. Instant PASS/WARN/FAIL against 20+ Australian government registers.

Start verifying →
VERIFY A SUPPLIER
Run a free check in seconds

Search by business name, ABN, or ACN. Get a real-time PASS/WARN/FAIL report across 8 verification checks.

Start verifying →

Contains data sourced from the Australian Business Register and ASIC, © Commonwealth of Australia, licensed under CC BY 3.0 AU.