NetSuite Next is live in NZ, but its agent controls haven't shipped

What Oracle released

New Zealand and Australian NetSuite customers can now move to NetSuite Next. Scott Wiltshire, Oracle NetSuite's ANZ vice president, called NetSuite Next the product's "biggest update since launching in 1998", IT Brief reported on 27 October. Ask Oracle is the conversational assistant that answers questions about a company's own NetSuite data. Oracle is rolling Ask Oracle out to existing accounts before those accounts switch to NetSuite Next. Back in March, Oracle had only committed to bringing NetSuite Next to ANZ within 12 months, so the regional rollout landed at the early end of that window.

Oracle announced the features themselves at SuiteWorld in Las Vegas on 26 October as part of what it calls an Autonomous Suite. IT Brief's write-up on 2 November set out the detail for ANZ readers. Oracle's release says four items are available now in every NetSuite Next market:

• SuiteSkills, more than 45 prebuilt AI functions
• AI Portlets
• AI Workspaces, which include an Autonomous Close workspace aimed at a zero-day month-end close
• a selection of industry SuiteAgents

According to IT Brief, the agents for software, product and construction businesses are live now. Healthcare and nonprofit agents are due in a later release.

Some of the most heavily promoted tools haven't shipped. Keith Kirkpatrick, an analyst at Futurum, points out that SuiteAgent Studio, SuiteFlow Agent and the SuiteAgent Management Workspace are missing from NetSuite's list of currently available capabilities. Oracle's own release says other Autonomous Suite capabilities are "planned for a future release". IT Brief adds that SuiteExpress, the faster implementation offering, is initially limited to software companies in the US and Canada.

Why the gap between features and controls matters

The tools now live in NZ go beyond reporting. IT Brief describes Ask Oracle's Model Context Protocol features as able to route approvals or place payments on hold, where permissions allow. An AI Workspace brings people, agents and approvals together on a single process such as the month-end close. Once software can act on transactions, the controls around that software carry far more weight than they did when NetSuite's AI only wrote summaries.

The central agent controls are the pieces still on the roadmap. Oracle describes the SuiteAgent Management Workspace as the place where administrators set agent access and execution permissions, watch activity, switch agents on or off, and trace record changes through System Notes. Kirkpatrick's assessment is that the controls that make autonomy safe are arriving after the autonomy itself. Until the workspace ships, NZ customers running the available agents will depend on NetSuite's existing role and permission model. Oracle's release confirms that SuiteSkills follow existing roles and permissions.

The practical consequence sits inside each customer's role design. A NetSuite account where broad administrator roles were handed out to save time during implementation will give an AI feature acting under one of those roles the same broad reach. Many organisations last reviewed their roles when they went live, and an AI rollout is a sensible prompt to review them again.

The zero-day close claim

Autonomous Close will catch the eye of NZ finance managers who lose a week each month to reconciliations. Kirkpatrick notes that Oracle describes zero-day close as a stated aim and cites no customer results, days-to-close reductions or referenceable case studies. He also argues that a credible zero-day close needs an audit trail auditors accept, plus finance leaders willing to sign off numbers assembled by agents. The customer savings in IT Brief's ANZ coverage, such as one retailer estimating a 50 percent cut in close time, are vendor-supplied case studies based on the customers' own estimates.

The sensible reading is that Autonomous Close is worth piloting on a single entity or a single reconciliation type. A company shouldn't rebuild its month-end calendar around Autonomous Close before its own auditors have seen the output.

The bring-your-own-model question

Customers can stick with NetSuite's embedded AI or connect an external large language model through the NetSuite AI Connector Service, which IT Brief reports more than 13,000 customers have adopted. External models give NZ businesses flexibility, and they also raise a privacy question. If customer, supplier or staff records go to a model hosted offshore, the business needs to know whether that transfer counts as a cross-border disclosure under Information Privacy Principle 12 of the Privacy Act 2020. IPP 12 allows such disclosures only where safeguards such as comparable overseas law or contractual protections apply. Whether IPP 12 applies depends on how the model provider handles the data, which is why the provider's terms deserve a careful read before anyone connects a model to the live ERP.

Kirkpatrick also points out that the same openness through MCP and the AI Connector Service lets agents from other vendors reach NetSuite data. For a NZ business already trialling assistants in Microsoft 365 or elsewhere, every connected agent becomes another identity with access to financial records and needs its own place in the permission model.

Where buyers are starting

Futurum's survey of 833 enterprise application decision makers, published in August, found that 62.5 percent plan to increase ERP spending over the next 12 months. The most common first areas for agent deployment were IT operations and observability, chosen first by 21.6 percent, followed by cybersecurity at 16.4 percent. Finance didn't lead that list, which matches Kirkpatrick's advice to begin with generally available tools in contained processes.

Next steps for NZ NetSuite customers

• Confirm what your account has. Check with your NetSuite partner or account manager whether you're on NetSuite Next or still using Ask Oracle within the existing release. Get a written list of which features are switched on.
• Review roles before enabling AI features. Remove unused administrator access and tighten roles that can approve payments or post journals. Make sure each AI feature runs under a role scoped to its task.
• Pilot one process with a clear measure. Pick a single reconciliation, an inventory exception report or AI Portlets on a management dashboard. Record the current time and error rate so the pilot produces evidence your team can trust.
• Check data handling before connecting external models. Read the model provider's terms on storage location, retention and training use. Then decide with your privacy officer whether IPP 12 applies.
• Hold the agent-heavy plans until the governance tools ship. Keep SuiteAgent Studio and wider agent rollouts on the roadmap until the SuiteAgent Management Workspace is available in your market, then reassess.
• Brief your auditors early. If Autonomous Close is on the table, ask your audit firm what evidence it would expect from agent-assembled reconciliations before you change the close process.

Getting value from NetSuite Next depends mostly on clean role design, sound data and processes that are documented well enough for an agent to follow. If you're weighing the upgrade alongside wider ERP and business systems work, or want to choose which finance processes are ready for automation, those foundations deserve attention before any agent is switched on.