By Manvinder Sidhu, Co-founder, Ezyiah
Short answer: MYOB AI BAS, launched in beta in March 2026, reviews transactions inside an MYOB file, suggests GST treatments and pre-populates a BAS for an accountant or bookkeeper to sign off. At launch it targeted simple, cash-based files. It can only work with data already in the MYOB file, so the expensive work of getting messy bank data in and correctly coded is still left to people. An AI bookkeeper such as Ezyiah takes on that work in its own ledger.
MYOB launched AI BAS in March 2026 as part of a suite of AI-powered features rolling out across its SME products, and MYOB describes it as Australia's first agentic BAS offering. In practice, the feature works inside the MYOB file: it reviews transactions, suggests GST treatments, flags anomalies for the business owner or adviser to resolve, and produces a pre-populated BAS report with lodgement totals for an accountant or bookkeeper to sign off before submission. It is a real step forward for the files it fits, and understanding exactly which files those are is the useful part, because the answer explains both what the tool does well and where the actual preparation work still lives.
What MYOB AI BAS does, and where its scope currently sits
The initial beta began with invited sole traders on Solo by MYOB and MYOB Business Lite and Pro, focused on simple BAS scenarios covering the core GST fields for cash-based accounts, with MYOB indicating at launch that access would broaden through 2026. Check MYOB's current release notes for where availability stands now. The positioning is deliberate and sensible: AI BAS prepares and pre-populates for adviser review rather than lodging on its own, the accountant or bookkeeper keeps the final sign-off, and MYOB is explicit that it is not itself a registered tax or BAS agent and that the user remains responsible for accuracy and compliance. For a straightforward sole trader file with clean bank feeds, well-coded transactions, and no catch-up periods, the feature genuinely reduces the manual work involved in getting to a reviewable BAS, and it keeps the file tidier through the quarter by surfacing missing documents and GST anomalies as it goes. That is a real benefit, and it is worth saying plainly before turning to the boundary.
The boundary is simply the current scope. The initial release targets simple BAS scenarios and cash-based accounts, so more complex arrangements, multi-entity structures, accrual-basis clients, and anything involving detailed apportionment or non-standard GST treatment sit outside it for now. More fundamentally, and this is inherent rather than a limitation of the beta, AI BAS can only work with what is already in the MYOB file. It cannot reach data that has not been imported. A client who sends a PDF bank statement that has not been converted and uploaded does not have that data in MYOB. A client whose CSV export failed on import because of formatting errors does not have those transactions in MYOB. A client with three months of unreconciled activity sitting in a spreadsheet because they never connected their feed does not have that data in MYOB. None of those situations is unusual. They are the standard reality for a large share of a mid-size firm's client book, particularly tradie clients, clients with multiple accounts, clients who send whatever format they happen to have, and clients working through catch-up periods.
The gap between the tool and the real workflow
This is not a criticism of what MYOB has built, it is a description of where the tool starts and where the time actually goes. For most firms the expensive step is not reviewing a pre-populated BAS, it is getting the data to the point where that review is even possible. That means converting PDF statements into an importable format, cleaning CSVs with date-format problems, missing columns or encoding errors, reconciling feed transactions against the manual imports used to plug gaps in the feed history, and correctly coding the overseas software subscriptions, the vehicle purchases under finance, and the mixed-use expenses before any automated review can trust the GST totals it is looking at. An AI BAS review that starts once the file is already clean leaves all of that with the accountant. The time only shrinks when the bookkeeping itself, not just the review at the end, is done by the software.
How Ezyiah handles the preparation
Ezyiah is an AI bookkeeper with its own double-entry ledger, so the preparation is part of the product, not something that has to be finished before it can start. It pulls live bank feeds daily and reads uploaded CSV and PDF statements, up to five at a time. It reconciles and categorises transactions with AI, remembers how you code each client with AI Memory, matches receipts, bills and invoices, and builds the cash-basis BAS, P&L, Balance Sheet and General Ledger from the same books. Your accountant reviews and approves before lodgement.
See Ezyiah vs MYOB, or how it works for accounting firms and business owners.
The takeaway
MYOB AI BAS is a genuine improvement for simple, clean, well-maintained MYOB files, and for those clients it cuts the manual work in BAS preparation and hands the accountant a cleaner starting point. The honest complication is that the clients whose files are already simple, clean and well-maintained are rarely the ones eating the most preparation time. The hours go to the clients whose data arrives in messy formats, who have missed periods, who bank with institutions without reliable feeds, and who send what they have rather than what the software needs. Those clients are not in AI BAS scope today, and more importantly, the part of their workload that is expensive, getting their data clean and complete in the first place, sits before the point where AI BAS begins. A firm evaluating any AI BAS tool, MYOB's included, is really asking two separate questions: how good is the review once the data is clean, and what gets the data clean to begin with. The tools are converging fast on the first. The second is where the real quarterly time still sits.
