A Bank Traced About $191,000 Into an Unaccredited Broker’s Account. The Tribunal Ruled Only on the Banker — the Referral Boundary Is What Brokers Should Read

What was actually decided — and what was not

On 2 November 2026, Deputy President Gerard Boyce of the Fair Work Commission dismissed an unfair dismissal claim brought by a former senior home lending executive at the Commonwealth Bank. The bank had dismissed him on 29 July 2025; he lodged his claim on 18 August 2025. Boyce found the dismissal “wholly justified and not unfair”, and assessed the applicant’s evidence as “wholly unreliable” — describing it, in a line that has travelled through the trade press this week, as “a chaotic collection of lies, nonsense, poorly matched ideas, and manufactured responses”.

Before going further, it is worth being precise about the limits of that decision, because the commentary around it has been looser than the record supports.

With that said, the detail is the point. Reporting by Australian Broker, Mortgage Professional Australia and The Adviser sets out a sequence that most brokers never get to see: how a single anonymous tip became a reconstructed referral network.

The evidence trail a lender can assemble

It started with an anonymous report to the bank’s internal whistleblower service on 19 November 2024, naming the executive and four other employees. A preliminary review cleared him. That is the first lesson, and it is not a comfortable one: the initial look found nothing.

What found something was widening the lens. The bank reviewed the loans the executive had refinanced across from his previous employer, and a shape emerged. A high proportion of the borrowers were self-employed. They had been introduced by two accredited referral partners. And many of them arrived with company financials prepared by the same accounting firm.

None of those three facts is wrong on its own. A broker or banker with a niche will naturally see clustered borrower types, repeat introducers and familiar accountants. The signal was the three appearing together, repeatedly, inside one person’s book. Across the two partners, reporting puts the volume at 23 customers, 35 applications and 28 funded loans; the larger of the two partners alone accounted for 17 customers and 29 applications between April 2024 and January 2025, 22 of which were funded.

From there the bank moved to transactional evidence, and traced money into the account of a broker who was not an accredited referrer under its program. Three streams were reported: six payments from the first referral partner totalling $81,146.44; six from the second totalling $61,441.62; and $48,595.50 from five of the referred customers themselves, moving around the time their loans were processed or approved. Transaction descriptions included “service fee” and “referral fee”. Together the three streams come to roughly $191,000 between August 2023 and February 2025.

Then came the device evidence. Asked at a 14 May 2025 interview who the broker was, the executive said he had “no idea”. Forensic analysis of his work phone found a saved contact whose number matched that broker’s own customer profile at the bank, along with messages and missed calls. One customer had texted to thank the executive and the broker together. A later explanation — that the man was someone he knew by a different surname — was rejected by Boyce as “unbridled nonsense” and a “complete fabrication”.

Finally, the channel itself. The executive admitted joining a group chat, named after the referral partner, that included the approved referrer and the unaccredited broker, and admitted that loan leads arrived there and that he reported back whether loans were approved or settled. He had deleted the chat and declined to hand over his personal phone. Boyce observed that what was actually discussed in it will never be known.

Where a referral stops being a referral

The structural issue underneath this case is the one brokers deal with every week: the line between a referral and credit assistance.

Under the National Consumer Credit Protection regime, a narrow licensing exemption lets a referrer do very little — in substance, tell a consumer that a licensee can help with credit and pass on how to contact them, while disclosing any benefit the referrer receives for doing so. The exemption is deliberately thin. Once a third party starts shaping which lender or which product the client should pursue, the activity looks like credit assistance, and credit assistance requires a licence or an appointment as a credit representative. The precise conditions are set out in the regulations and are easy to breach by accident, so the sensible step is to have your licensee or aggregator’s compliance team confirm how each of your arrangements is documented rather than relying on how it has always worked.

Lenders police their side of that line through their own program rules. In this case, the bank’s referral program reportedly prohibits mortgage brokers and builders from participating directly or indirectly, restricts contact with the referrer during assessment to the bank’s own business development staff, and prohibits sharing customer information with referrers. The group chat cut across all three.

For brokers, the practical read-across is about introducers flowing the other way. If a referrer sends you clients and is also being paid by those clients, or is also preparing their financials, or is also talking to a lender about their files, your arrangement has more moving parts than your referral agreement probably describes.

The two admissions that matter most to a broker

The misleading-the-investigator finding was the contested one. But the two allegations the executive admitted are the ones that map most directly onto broker work.

1. Income that was not verified because verifying it was hard

He left PAYG employment income out of two applications and had the borrowers assessed as self-employed, even though their bank statements showed regular wage credits. He first told investigators he had forgotten. He later said he had been “lazy to verify the PAYG income” because it was “hard to verify”.

Set aside the candour of the phrasing. The mechanism is ordinary: an income stream that complicates a file gets left out, and a servicing position that would not otherwise work does. Any broker who has felt the pull of a clean self-employed assessment over a messy mixed-income one knows the shape of this. The borrower’s own statements were the thing that gave it away — which is to say, the document already on the file.

2. Customer information moving through a consumer chat app

Sharing loan progress with a referrer and a third party through a group chat was found to breach the bank’s technology policy. Brokers are not bound by that policy, but the privacy and record-keeping exposure is comparable, and worse in one respect: a broker who conducts client business in a chat app has created file records that sit outside the CRM and may not be retrievable when an audit, a complaint or a licensee review asks for the complete file. Deleting the chat did not help here. It left a gap that the decision-maker simply read against the person who created it.

Why this lands now

The timing is what lifts this above a one-off employment dispute.

ASIC has flagged a review of lender conduct within its banking supervisory agenda, expected to begin in the third quarter of 2026–27, covering three things in particular: the effect of changes banks made to proprietary lender incentives, lenders’ use of referrers and the controls around third-party introducer arrangements, and how lenders monitor broker-originated lending — including accreditation and re-accreditation, application quality, fraud indicators and customer outcomes.

That review will be conducted against a backdrop brokers already know. The Commonwealth Bank self-reported suspected fraudulent home lending to authorities in February 2026, with reporting putting the figure at around $1 billion and industry-wide estimates since reported above $3 billion. AUSTRAC’s Fintel Alliance work found the same intermediaries recurring across suspect applications at multiple banks. The MFAA has been developing referrer risk standards; its chief executive, Anja Pannek, has made the point that fraud does not respect organisational or distribution boundaries. And ASIC’s Report 840, published on 13 October 2026, found gaps in how brokers demonstrate that their recommendations serve the customer’s best interests, particularly where a cheaper loan was available.

Put those together and the direction is clear enough. Lender oversight of the third-party channel is going to be examined using exactly the artefacts this decision describes.

Your referral source audit: six checks worth running this week

What to watch next

Three things are worth tracking into 2027. The first is the scope of ASIC’s lender-conduct review when it formally begins, and in particular whether its broker-oversight limb results in changes to accreditation and re-accreditation practice at individual lenders — the kind of change that arrives as a panel notice rather than a regulatory announcement.

The second is introducer programs themselves. Reporting notes that two of the major banks continue to operate them while two closed theirs years ago, and that industry voices have called for them to be scrapped entirely. If a bank withdraws from introducer referrals, the referral flow does not disappear; it looks for another channel, and brokers are the obvious one. That is an opportunity, and it comes with the obligation to document the arrangement properly from day one.

The third is the quiet one: file completeness. The case turned substantially on records that existed outside the official system — a phone contact list, a deleted chat, a payment trail. Brokers who can produce a complete, coherent file from a single system are in a materially different position from brokers who cannot, and that gap is widening as reviewers get better at asking.

None of this requires a broker to treat good referral partners with suspicion. It requires the arrangement to be written down accurately, the money to be visible, the documents to come from their source, and the conversation to happen somewhere it can be retrieved. That is a modest amount of work set against the cost of explaining, later, a relationship that was only ever informal.