What Australia released
On 12 November 2026, Australia's Minister for Employment and Workplace Relations, Amanda Rishworth, released a draft Best Practice Guide on the Adoption of AI in the Workplace. The draft sits under three themes called Plan, Enable and Embed, and sets out eight principles. These run from identifying use cases and assessing workplace readiness through to establishing AI governance and proactively managing risks and work health and safety, according to On Notice's summary of the draft. Public comment is open through the Department of Employment and Workplace Relations consultation hub until 27 January 2027.
AAP describes the guidelines as voluntary. The department built the draft from roundtables with unions, employer peak bodies, academics and small businesses. Microsoft, Commonwealth Bank, Telstra, Kmart, Blackmores and Ramsay Health Care were the large employers named as contributors. The draft also carries case studies. In one, a Blackmores manual data-entry task fell from three days to 10 minutes.
Why the release drew political heat
The guide landed two days after Qantas said an offshoring deal with Accenture would send 850 Australian jobs overseas. The Australian Services Union called the Qantas decision AI-motivated, and Rishworth said Qantas should have offered the affected workers other roles, as the guidelines describe. Employer groups were less enthusiastic. Andrew McKellar, chief executive of the Australian Chamber of Commerce and Industry, said the draft's repeated calls for engagement before a decision sit uneasily with running a business. He argued that consultation out of proportion to the risk and scale of a change would slow adoption and add cost, and warned against AI standards becoming a back door to heavier regulation.
Both sides of that argument are worth hearing. Rishworth's own summary of the guide's position is short: "The Best Practice Guide makes it clear that AI works best when the workforce is at the centre of the adoption process." She also told businesses to work out what they want AI to achieve before telling staff to use it, and she warned that poor "bolt-on" adoption could harm the businesses doing it.
How the guide lines up with NZ law
New Zealand has no equivalent workplace AI guide, and no AI-specific statute either. NZ employers already work under rules that cover much of what the Australian draft recommends. Employment New Zealand counts changes to an employee's duties as workplace change, alongside disestablishing, merging or creating roles. For that kind of change, employers have to consult affected staff. That means giving them the reason for the proposal and reasonable time to respond, including time to suggest alternatives, and then considering their feedback before deciding. The same page lists improved technology as one example of a genuine business reason for change.
The practical effect is that an AI rollout which reshapes an accounts payable clerk's daily work can trigger a consultation process even when nobody's role is at risk. Many NZ firms treat AI tools as software purchases handled by IT. Employment law treats a material change to duties as an employment matter, whatever caused the change.
Health and safety law adds a second obligation. WorkSafe's guidance says businesses must engage workers directly affected by health and safety matters, including when proposing changes which may affect workers' health and safety. An AI tool that changes work pace, monitors staff activity or reallocates tasks may meet that test. The answer depends on what the tool does in your workplace, so the assessment needs a person who understands both the system and the work.
The Australian draft gives NZ employers a ready-made structure for meeting obligations they already have. The eight principles read like an outline for an internal AI adoption policy, and the "Establishing AI governance" principle covers ground that many mid-sized firms haven't documented yet. For trans-Tasman businesses there's a more direct reason to read the draft: Australian staff, unions and customers will start treating it as the benchmark for reasonable conduct, even though it's voluntary.
Where we'd apply judgement
McKellar's point about proportionality deserves a fair hearing. A pilot that helps staff draft emails doesn't need the same process as a project that automates most of a back-office team's invoice matching. Running full consultation for every low-impact tool wastes staff goodwill and slows useful work. Skipping consultation on a high-impact change exposes the business to good faith claims and lost trust. A tiered approach handles both. Classify each AI use case by how much it changes duties, workload, monitoring or headcount, and scale the process to match.
The final guide hasn't been published, and nobody knows yet how much the draft will change after consultation closes in January. The NZ obligations described above don't depend on Australia's final version.
Practical next steps
- List your AI tools. Cover tools in use, in pilot and planned, and mark which ones change duties, workload, staff monitoring or headcount. Include tools that staff have bought themselves.
- Consult on duty changes. For any use case that changes duties, follow the Employment New Zealand change process before the final decision: a written proposal, the business reason, time to comment and a documented response to feedback.
- Involve health and safety representatives. Bring them in when a tool changes how work is paced, allocated or monitored, as WorkSafe's engagement guidance expects for changes affecting health and safety.
- Check your policy against the eight principles. Use them as a gap check on your AI governance policy. Name an owner for each AI system and set out who approves new use cases. Our AI consultancy team works on this kind of governance framework.
- Measure before rollout. Record how long target tasks take today, so any productivity claim you make to staff or the board can be checked. Our process and automation work starts from the same baseline.
- Trans-Tasman firms: consider a submission. If you employ people in Australia, read the draft and consider making a submission through the DEWR consultation hub before 27 January 2027.
Sources
- Draft best practice guide for AI adoption in the workplace released (National Tribune)
- First workplace artificial intelligence guide released (AAP)
- Draft AI workplace guide, built from roundtables with Microsoft, CBA, Telstra and Kmart, sets 8 principles for adoption (On Notice)
- The basics of workplace changes (Employment New Zealand)
- What is worker engagement and participation? (WorkSafe New Zealand)
Want to know what this means for your organisation? Our team in Auckland can show you your top three opportunities in a free 30 minute assessment.
Get my free assessment How we help