MYOB Acumatica AI: What Australian Finance Teams Can Switch On Today

Every few weeks another AI announcement lands from Acumatica, MYOB or Microsoft, and somewhere in Australia a finance manager opens their MYOB Acumatica tenant looking for the feature it described. Usually it isn’t there yet. Sometimes the announcement covered the global Acumatica product, which runs months ahead of the ANZ release, and the feature won’t arrive for two or three quarters.

That gap causes real planning mistakes. So here’s the sorted version: what MYOB Acumatica AI can do in the release Australian customers are actually running, what’s in preview, and what’s still a promise with a date attached. Everything below traces to a vendor source checked this month.

What MYOB Acumatica AI includes today

MYOB Acumatica AI currently means two shipped features plus one preview. The shipped features arrived in the 25 R2 release, MYOB’s first major MYOB Acumatica release of 2026.

The Cross-Sell Assistant analyses your historical sales data and suggests complementary products at the point of sale, so the cross-sell relationships don’t need to be maintained by hand. For businesses on the Manufacturing Edition, Production Variance Anomaly Detection watches production costs, labour time and efficiency, and flags unusual movements so teams can investigate early instead of finding the problem in month-end reporting.

Notice what both features have in common: they suggest and they flag. Neither posts a transaction, changes a price or touches the ledger. A salesperson decides whether to offer the suggested product. A production manager decides whether the flagged variance matters. That’s the shape of AI we tell clients to trust first, and it’s the same drafts-first pattern we apply to AI agents over ERP data generally.

Demand for this is not speculative. MYOB’s ERP Trends research found 78% of wholesale distribution businesses plan to use AI in their ERP, and the ABS reports that 12% of Australian businesses used AI in 2024-25, rising to 35% of large businesses. The question for most mid-market finance teams is no longer whether to use AI, it’s which features are real in their release.

AI Studio: describe an agent, get a draft

The most interesting piece of 25 R2 is AI Studio, and it’s important to label it correctly: tech preview now, with general availability planned for 26 R1. Existing customers can register interest with MYOB to join the preview.

AI Studio lets you build small AI agents inside MYOB Acumatica using plain-language instructions. Describe the task, connect the agent to the relevant business data, and it generates the output inside the platform. MYOB’s launch examples include drafting product descriptions for large catalogues, summarising supplier transaction history and preparing customer service responses.

Here’s how that looks as a worked pattern for a wholesale distributor. An agent reads your open purchase orders and each supplier’s delivery history, identifies the orders most likely to slip based on that history, and drafts a chasing email for each one with the PO details filled in. The purchasing officer reads the drafts each morning, edits the two that misread the situation, and sends the rest. Nothing leaves the building without a person deciding it should. The agent removes the hour of looking things up, not the judgement.

The architecture matters for Australian buyers. MYOB states that AI Studio is powered by AWS and Anthropic, that AI processing runs in local data centres so data never leaves Australian or New Zealand borders, and that your data is never used to train external models. If your business already has a preferred AI provider, you can connect that instead. Those are exactly the questions a CFO should ask of any AI feature, and it’s notable that MYOB answered them in the launch material rather than in fine print.

A preview is still a preview. Capabilities can change before general availability, and you shouldn’t build a month-end process around a feature that’s still being shaped. Use the preview window to learn where your data lets you down.

Why Acumatica’s AI headlines aren’t in your system yet

This is the part most coverage skips. Acumatica, the global platform MYOB Acumatica is built on, released its 2026 R1 update in March 2026, including an AI Assistant that’s rolling out under a managed availability program rather than as a general release. Global Acumatica announcements generate most of the AI headlines you’ll read.

MYOB Acumatica typically receives global Acumatica features 6 to 9 months after the global release, because MYOB localises each release for ANZ payroll, GST and compliance requirements before shipping it. The current MYOB Acumatica release is 25 R2. The global 2026 R1 features are not in it, and MYOB hasn’t published a committed ANZ date for them.

So when you read about an Acumatica AI capability, ask one question before it goes anywhere near a plan: which release is that in, and is it a MYOB release? Your partner should be able to answer from the release notes. If they can’t, that tells you something too.

What the Microsoft partnership promises, and when

In April 2026, MYOB and Microsoft signed a five-year partnership to jointly fund and build AI features across MYOB’s products, using Microsoft’s Foundry, Copilot Studio and Agent 365 tooling. For MYOB Acumatica specifically, the announcement promises native AI features for contextual financial insights, natural-language queries and AI-assisted document processing.

All of that is roadmap. The first jointly developed customer features are due later in 2026, and none of the promised MYOB Acumatica capabilities have shipped as of August 2026. The partnership is a credible signal of where the platform is heading, and the natural-language query direction lines up with what our clients keep asking for. But a signed partnership is not a feature in your tenant. Don’t build a business case on it until release notes say otherwise.

Switching things on safely

The features above are the vendor’s side of the story. The other side is yours: an AI feature is only as safe as the controls around it and only as useful as the data underneath it.

The controls are the same ones we’ve written about for AI agent governance: role-scoped access so an agent sees only what the requesting user could see, human approval on anything that writes to the ledger or leaves the business, and an audit trail of what was suggested and who accepted it. If customer or employee data flows through an AI feature, check your obligations under the Australian Privacy Principles via the OAIC’s guidance on commercial AI products before you switch it on.

The data question decides more outcomes than the model does.

“The businesses getting value out of these previews aren’t the ones with the most ambitious use cases. They’re the ones with clean item files, tidy supplier records and consistent naming. Point an agent at messy data and you get confident drafts of the wrong thing.”

Bill Alvarez, Practice Manager, Auboros

If you want to be ready rather than waiting, the useful work this quarter is unglamorous: clean the master data the first agents will read, decide which workflows you’d trust to drafts-first automation, register for the AI Studio preview, and put a one-page AI usage policy in front of your team. We covered what that looks like in practice in our guide to what AI agents actually do inside Australian businesses, and the reconciliation agent pattern shows the approval-gate design end to end.

Working out what MYOB Acumatica AI is worth to your business?

Auboros implements and supports MYOB Acumatica across Queensland and beyond, and we run AI agents against live ERP systems in our own operations, so we can tell you which of these features will earn their keep in yours. Book a free consultation and we’ll go through it with your data, not a demo script. No pressure either way.