Bunnings cuts, TechnologyOne holds staff: two AI workforce plans

By DataForge team18 November 20265 min read

On 17 November Bunnings cut 350 tech roles and TechnologyOne said it'll double R&D output with flat headcount. What NZ firms should decide first.

Two plans from the same Tuesday

On 17 November 2026, two large Australian companies announced opposite plans for how AI will change their technology workforces.

Bunnings will cut about 350 technology roles. That's 230 permanent positions plus roughly 120 contractor engagements, and the work moves to Tata Consultancy Services teams in India and the Philippines under a multi-year partnership (Inside Retail). About 500 technology staff will stay in-house in Australia. The handover starts in December 2026 and runs to June 2027. The Nightly reported that Bunnings tied the decision to keeping pace with the "artificial intelligence revolution". Yahoo Finance put the reduction at around 40 per cent of the retailer's internal technology staff. Bunnings hasn't published the savings it expects, and none of the reports say whether its New Zealand technology roles are affected.

The same day, TechnologyOne, the Brisbane ERP vendor, reported its year to 30 September 2026. Profit before tax rose 20% and annual recurring revenue rose 18% to $652.1 million (TechnologyOne). Chief executive Edward Chung told Reuters the company expects to "probably double the output" of its research and development staff with minimal headcount growth. The company now employs more than 1,600 people, up from more than 1,500 a year earlier (Reuters, via 104.9 The Wolf). On the results call, Chung ruled out mass layoffs and said the company plans to keep headcount reasonably flat while meeting its growth targets (Investing.com transcript).

What separates the two approaches

Both companies want more output from each dollar spent on technology. Bunnings is getting there by shrinking its in-house team and buying delivery capacity from an outsourcer. TechnologyOne is keeping its engineers and expecting AI tools to raise what each person produces. The Bunnings model turns a fixed salary cost into a contract cost. The TechnologyOne model keeps institutional knowledge in-house but depends on productivity gains actually showing up.

TechnologyOne has published more detail, and its figures show how early the evidence still is. Total costs grew 8% while recurring revenue grew 18%, according to the results call. R&D spending still rose 17% to $179.9 million, which is 26% of revenue (IT Brief). Management also said the share of R&D it capitalises rose two points to 57%, and put the whole increase down to AI development. Finance leaders should notice that detail. A higher capitalisation rate improves reported profit in the current year, so part of the profit gain is an accounting effect. It doesn't show that work got cheaper.

Bunnings' AI case rests on its partner, and Tata's delivery model is also unproven at the retailer's scale. Inside Retail didn't mention AI at all and described the deal as cost reduction plus added capability. Bunnings is presenting the change as AI-driven, but the published facts look like a classic offshoring contract.

The pricing detail NZ customers should read

New Zealand makes up 7% of TechnologyOne's business, according to the results call. Its main customers are councils and tertiary providers. The company charges a flat 30 Australian cents per AI interaction so customers aren't surprised by computing costs (Reuters, via 104.9 The Wolf). Management said usage of conversations and interactions is well above expectations. It plans to sell its Plus and Guide AI products as bundles of conversations that count towards recurring revenue.

A vendor that's pleased usage is beating forecasts is also telling customers their bills will grow faster than forecast. A council or university that budgeted AI as a fixed add-on will need to track usage per user, per team and per process. Paying 30 cents per interaction looks trivial until a staff-facing assistant is answering thousands of routine questions a week.

What this means for NZ businesses

Most New Zealand organisations will face the same choice on a smaller scale. A 40-person finance and operations team can't go to a global outsourcer the way Bunnings did. It can still choose between holding headcount steady while expecting more output and letting attrition shrink the team as AI handles routine work. The wider Australian picture shows how often employers are taking the second option. Yahoo Finance listed AI-linked cuts at Qantas (850 roles), Canva (1,271), NAB (2,500) and ANZ (4,500). The same report said the Reserve Bank of Australia has seen only tentative signs of adjustment in the labour market.

We think the TechnologyOne approach suits most mid-sized NZ firms better, because small teams lose a lot when one person who understands a process leaves. Holding staff only pays off if someone measures the gain. TechnologyOne can point to recurring revenue growing faster than costs. Most businesses have no comparable baseline for how long invoice processing, month-end close or service desk triage took before AI tools arrived. Without that baseline, any productivity claim is a guess, whether it comes from staff or from a vendor.

Outsourcing to an AI-enabled provider deserves the same scrutiny. If a provider's pitch depends on AI lowering delivery costs, the contract should show where those savings land, and it should say who owns the prompts, workflows and process documentation once the provider has built them.

Practical next steps

  • Write down your workforce position before buying more AI tools. Decide whether you're aiming for more output from the same team or the same output from a smaller one. Both are legitimate choices. Staff and vendors behave differently depending on which one you pick.
  • Record baselines now. Pick three or four processes, such as accounts payable, payroll queries or IT ticket triage, and log current volumes and cycle times. Our AI consultancy work starts with this step because it's the only way to tell a real gain from a felt one.
  • Read the AI pricing clauses in your ERP and SaaS renewals. Look for per-interaction, per-conversation or credit-based charges. Ask the vendor for usage reports at team level, and set an alert threshold before the first bill.
  • Check capitalisation and cost treatment with your finance lead. If AI-assisted development or configuration work is being capitalised, make sure reported savings reflect real cash rather than accounting treatment.
  • If you're considering outsourcing, put AI savings into the contract. Specify productivity commitments, rate reductions over time, and ownership of any automations built on your data. Our process and automation team can help map which workflows are worth keeping in-house.
  • Talk to staff early. Whichever model you pick, tell affected teams what AI will change in their roles and what it won't, before tools go live.

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
Get new articles by RSS