Qantas's AI overhaul starts with 1,000 legacy apps: lessons for NZ

What Qantas announced

On 10 November, Qantas announced a multi-year partnership with Accenture that the airline expects to affect about 850 Australian roles over the coming years (iTnews). Reuters puts the cut at roughly 3% of Qantas's Australian workforce of 28,000. Finance, procurement, human resources and marketing are in the first phase, and the first round of changes is due by mid-2027 (Reuters via WBOW). Accenture teams in India will deliver most of the transferred work, some will stay in Australia, and the programme runs internally under the name Project iQ (ACS Information Age).

Qantas has presented the deal as a response to rapid, AI-driven change in aviation. Chief executive Vanessa Hudson told staff the airline can reach data and AI skills through a partner faster than it could hire them (Reuters via WBOW). The technical detail says more about the plan. iTnews reports that Qantas wants to deal with decades of technical debt spread across more than 1,000 legacy applications that were never designed to work together. Accenture will invest upfront to speed up modernisation of the airline's data platforms. Qantas's own statement calls connecting data across the modernised systems "the critical next step" toward faster AI adoption (iTnews).

Why the AI label is contested

The Australian Services Union accused Qantas of "using AI to launder their cuts to Australian jobs". ACTU secretary Melissa Donnelly said unions worry about companies using AI as cover for job cuts with no accountability (Reuters via WBOW). In the same Reuters report, University of Sydney Business School professor Uri Gal said the Qantas decision looked like conventional outsourcing. Reuters also named WiseTech, Atlassian and Telstra as Australian companies facing "AI washing" claims. The Reserve Bank of Australia has said AI hasn't yet caused widespread job losses, though job ads are weakening in occupations exposed to AI.

The two sides also dispute what was said earlier. The ASU says Qantas told the union in August there were no plans to offshore jobs under the Accenture deal (ACS Information Age). A Qantas spokesperson denied ever giving that assurance and said the airline had met the union several times about moving roles to Accenture (Reuters via WBOW).

We think both readings fit the published facts. The near-term change is a sourcing decision combined with a large systems and data clean-up. The AI layer comes later, since iTnews reports that services moving to Accenture will in time be transformed further with automation and AI (iTnews). Qantas's public statements don't say how the 850 roles split between offshoring, redesign and automation, and they don't name the AI tools the airline plans to use.

What the deal means for NZ businesses

Few New Zealand organisations run anything close to 1,000 applications. The underlying problem still shows up at much smaller scale. Picture a mid-sized distributor running an ageing ERP, a separate warehouse system, a CRM, pricing in spreadsheets and a dozen integrations added over 15 years. Any AI tool placed over an estate like that inherits every mismatch in customer codes, product data and approval rules. Qantas spends more than $1 billion a year on technology (ACS Information Age), and even so the airline decided its systems and data had to be connected before AI could do much with them. For a NZ firm planning an AI programme, that sequencing is the most useful part of the announcement. Our digital transformation work starts from the same point, because AI pilots built on fragmented data tend to stall once they leave the demo stage.

Hudson's argument about hiring will sound familiar to plenty of NZ operations and IT leaders, who struggle to recruit data engineers and integration specialists at the speed projects need. A partner can supply delivery capacity. The client organisation still has to own process design, data definitions and the decisions about which work gets automated, because a partner can't make those calls without knowing how the business runs. Qantas is keeping an onshore AI capability at its Adelaide innovation centre, which iTnews says will eventually house about 400 staff (iTnews). A NZ business that outsources development or support should keep at least one person in-house who understands the end-to-end process and can challenge what the partner proposes.

The backlash also shows how quickly an AI explanation gets tested when roles are affected. Reuters reported scrutiny from unions, academics and a Greens senator within two days of the announcement (Reuters via WBOW). NZ employers proposing restructures will find consultation easier when the proposal names the actual drivers, whether those are sourcing, system replacement or automation of particular tasks. Staff, customers and the media respond better to a specific case than to a broad appeal to AI, and a specific case also gives the business a baseline for checking later whether the change delivered what it promised.

Australian workplace debates often reach New Zealand within a year or two, through shared employers such as banks, telcos and retailers. NZ subsidiaries of Australian groups should expect head offices to look at similar partner models for back-office work. Local leaders will want their own view of which processes and data are ready before group decisions land.

Practical next steps

• List your business applications and integrations, and record which system holds the master record for customers, products, suppliers and staff.
• Choose one back-office process, such as accounts payable, procurement or staff onboarding, and map the process end to end before you choose an AI tool for it.
• Estimate the data clean-up needed before an AI pilot can run on real records, and put that cost in the business case.
• If you're bringing in a delivery partner, write down which process knowledge, data ownership and decision rights stay inside the business.
• Capture baseline measures now, such as cycle times, error rates and rework volumes, so you can test later whether automation changed them.
• When a change affects roles, describe the real drivers in the consultation proposal and avoid putting it all down to AI.
• Agree who approves new AI uses and how they're reviewed. Our AI consultancy page covers adoption and governance planning for NZ and Australian organisations.