A common shape for small Australian tax practices: one senior registered agent (the principal), one or two unregistered staff doing the bulk of preparation work, and an assumption that because there's a registered agent on the letterhead, supervisory arrangements are covered.
Since TPB(I) 48/2024 came out, that assumption is on shakier ground than it used to be. The supervision obligation has always existed under the Tax Agent Services Act 2009, but the 2024 guidance is more explicit about what the TPB expects to see when it asks.
What "supervision" means in TPB language
Supervisory arrangements - in TPB(I) 48/2024's framing - are the arrangements for directing, overseeing and checking the tax agent or BAS services performed on behalf of a registered practitioner, to ensure they're provided competently. The key word is checking. Direction and oversight aren't enough on their own; there has to be a review function with a paper trail.
In essence, the TPB is asking three things:
- Is there a sufficient number of registered individuals for the services being provided
- Are the supervisory arrangements adequate in scope and rigour for the work being done
- Can you show evidence that the arrangements are operating
There's no fixed formula for the first one. A practice with three agents and twenty staff doing high-volume BAS work is in a different position to a sole practitioner with one part-time assistant. The TPB's assessment is case-by-case and turns on the actual workload, complexity, and risk profile.
The factors TPB(I) 48/2024 lists
The guidance enumerates the factors the TPB weighs when assessing whether supervision is adequate. Worth knowing what's on this list because it's broader than most firms expect:
- Financial, technological and human resources - do you have what you need
- Size and scale of services - volume of returns, BAS, advice
- Complexity of services - SMSF, GST registrations, ATO disputes are different to standard returns
- Staff qualifications and training frequency - not just initial qualifications, but ongoing
- Technology support and digital monitoring processes - how systems support oversight
- Quality assurance and control practices - file reviews, sample checks, peer review
- Risk management processes - how you identify and respond to issues
And separately, for any remote or hybrid working arrangement:
- Frequency of contact between supervisor and supervisee
- Supervisor availability during working hours
- Access to training resources
- Workflow management - how work is allocated and tracked
- Document review procedures - the part where the supervisor actually checks
- Remote audit and review capability - whether your systems support someone (you, or TPB) looking back at what happened
- Whether supervisors oversee multiple entities - splitting attention across practices is a flag
This second list became more prominent after the post-COVID shift to remote and hybrid practices. If your supervision model assumes everyone's in the office and they're mostly not, that's worth a second look.
Supervisory plans - recommended but not sufficient on their own
The TPB recommends putting a supervisory plan in place - a document that sets out who supervises whom, what's reviewed, how often, and what gets escalated. It's a good thing to have. But the guidance is also clear that a plan on its own doesn't satisfy the obligation. Whether the plan's adequate depends on whether it actually operates, and whether there's evidence of it operating.
In other words: writing the plan is the easy bit. The hard bit is the file review notes, the training records, the workflow audit trails that show the plan was followed last quarter and the quarter before that.
Where small practices typically come unstuck
A few patterns I see fairly often:
- The "I check everything before it goes out" model with no record of what was checked or when - works in spirit, fails on evidence
- Verbal training that never makes it into a training record - the training happened, it's just not visible to a reviewer
- Workflow that lives in someone's head rather than in a system - which means if that someone leaves, the supervision trail goes with them
- Reliance on the practice management system "noting" who logged the change without an actual review checkpoint
None of these are catastrophic on their own. Stacked together they're the difference between a practice that can answer the TPB's questions in an hour and one that spends a fortnight reconstructing records.
So where to from here
If you're a sole practitioner with no employees doing tax work, most of this doesn't apply directly - the supervision obligation is about overseeing other people's work on your behalf. If you have any unregistered staff doing prep work, the obligation is live and the question is how visible your supervision is.
A workable starting point is to write down what review actually happens today, however informal. Then look at it and ask: if a TPB compliance officer asked for evidence of these reviews over the last six months, could I produce it? If yes, you're in good shape. If not, the gap is where the work is.
myQMS handles a fair bit of this as a by-product of workflow assignments - work moves through stages, each stage captures who did what and when, and the supervision trail is the natural output. But the broader point holds regardless of tooling: supervision that doesn't leave evidence is supervision that's hard to defend.
Always happy to talk through what your current setup looks like if you'd like to compare notes.