AI Labor Tracker

Qantas Says Its Accenture Deal Is About AI. The Work in 850 Roles Moves Mostly to People in India First.

Qantas announced a multi-year Accenture partnership that will affect about 850 corporate roles, starting with finance, procurement, people and marketing. Read the order of the sentences: the work moves to Accenture teams, mostly in India, and only later gets automated. That order is the oldest play in corporate services, and it tells you what's paying for the deal.

On Tuesday in Sydney, which was Monday evening in New York, Qantas Group chief executive Vanessa Hudson sent an all-staff email. “Today we’re announcing a new multi-year partnership with Accenture and changes that will affect parts of our corporate functions,” it began, according to Australian Aviation, which saw the memo. “I know this is hard news for some, and I’m sorry for the uncertainty it brings.”

The airline’s public statement carried the headline “Qantas Group fast-tracks technology and AI investment with new Accenture partnership.” About 850 roles are expected to be affected over the coming years, with the first wave of changes in place by mid-2027.

The Australian Services Union said it believes most of the affected jobs are at the Qantas head office in Mascot, Sydney. Its New South Wales secretary, Angus McFarland, spent the day there meeting members. “This is jobs offshore and this is skills offshore at our national airline,” he told reporters.

Qantas shares rose more than 6% that day, ABC News reported.

So which is it: an AI story, or an offshoring story? The answer is in one sentence of Qantas’s own statement, and I think it’s worth reading slowly.

Read the sentence in order

Here’s the sentence, from the Qantas statement:

“Some of the services they provide will move to Accenture, delivered mostly by its teams in India and some in Australia, and in time be transformed further with automation and AI.”

There are three steps in that sentence, and they come in a specific order.

First, the work moves to Accenture. Second, it’s delivered mostly by people in India. Third, “in time,” it gets automated.

Step two is people. Accenture’s teams in India are human beings doing finance, procurement, HR and marketing work for an Australian airline. On day one of this deal, the work isn’t done by AI. It’s done by different people, in a different country, on a different payroll.

Hudson’s memo to staff, as quoted by Nine, says the same thing: “The first phase will start with finance, procurement, people and marketing, with some of the services in these areas moving to Accenture, delivered mostly by its teams in India and some in Australia.”

So the honest headline for mid-2027 is: some Qantas corporate work moves from Sydney to India. The AI part comes later, and its size and timing aren’t specified.

Why the order matters

Think about who pays for a deal like this, and how.

The Qantas statement says Accenture “will invest upfront to accelerate the modernisation of the Group’s data platforms.” A vendor doesn’t invest upfront out of generosity. It invests because it expects to earn that money back over a multi-year contract. The question is where the margin comes from in the early years, before any of the AI benefits show up.

The simplest answer is the oldest one in corporate services: labor cost. Moving finance, procurement and HR processing to a large delivery center in a lower-wage country has been the core of the business process outsourcing industry for decades. It works, which is why it’s still around. A company signs a multi-year deal, the vendor takes over the work, the work gets done at a lower cost per hour, and part of the difference funds the vendor’s investment and margin.

I can’t tell you the contract value or the cost per role in this deal. Neither Qantas nor Accenture has published them, and I’m not going to guess. But the structure Qantas described is the standard one. Offshore first, transform later.

There’s a reason the order runs that way, and it isn’t just money. To hand a process to a vendor, you have to write it down: every step, every exception, every approval, every system it touches. Accountants call it process documentation. Outsourcers call it knowledge transfer. Once the work is written down and running in one place, at volume, it’s far easier to automate than when it lived in the heads of a few hundred people spread across a head office.

That’s the first-principles case for why “offshore first, AI later” isn’t a contradiction. Codifying the work is the expensive part of automating it. Offshoring forces the codification and pays for it with wage savings. Then the automation gets layered on top, by the vendor, on a process it now runs.

If that’s right, it tells you something uncomfortable about the long run for these roles. The Sydney jobs go first, to people in India. Some of those Indian roles may later shrink as automation is layered in. The statement doesn’t promise either outcome, but “in time be transformed further with automation and AI” is exactly where that would happen.

What Qantas is betting on, in its own words

To be fair to Qantas, the AI language isn’t empty. The statement gives a concrete reason for the deal, and it’s a good one.

“Like every airline, Qantas runs on technology built up over decades for a different era, including more than 1,000 legacy applications that were never designed to work together,” the statement says. Modernizing those systems and connecting the data “is the critical next step to remove complexity, unlock innovation and accelerate the Group’s use of AI.”

That’s a real constraint. AI tools need clean, connected data to do anything useful in a company’s back office. An airline with over 1,000 applications that don’t talk to each other can’t simply turn on an AI agent and have it reconcile invoices. Somebody has to do years of plumbing first.

Qantas says it already spends more than $1 billion a year on technology, that thousands of employees already use AI tools in their daily work, and that its Product Innovation Centre in Adelaide is set to grow to more than 400 roles. It also says it’s a net employer: its Australian workforce is 28,000 strong, and it has created more than 4,400 new operational roles since 2023, including pilots, cabin crew and engineers.

Those numbers are worth holding next to the 850. Qantas is growing the jobs that fly and fix aircraft and shrinking the jobs that process paperwork about them. That’s consistent with what I’d expect from first principles: physical, safety-critical, on-site work is hard to move and hard to automate, while back-office work that can be written down is easy to move and, eventually, easier to automate.

“We need to reinvent the processes, platforms and data behind our business so we can innovate faster,” Hudson said in the statement. The goal, she said, is redesigning how the work gets done “with automation and AI built in from the start.”

The union’s question

The union’s response focused on something the statement didn’t address: what Qantas said earlier.

Reports of the Accenture deal, known internally as Project iQ, first emerged in August, Nine reported. At the time, ASU assistant national secretary Scott Cowen said the union had been told “there is no plan to offshore any Australian jobs.” On Tuesday, Cowen said workers deserved an honest explanation of what had changed since then.

“Earlier this year, Qantas told us directly there was no plan to offshore any jobs,” he said, according to Nine. “Today, they’ve come out and told hundreds of staff members that their work will now be delivered mostly from people overseas.”

I can’t independently verify what was said in a private meeting in August. What I can say is that this is the cost of framing an offshoring deal as an AI deal. When a company talks mainly about technology and AI, workers hear “our jobs will change.” When the details arrive and the work is going to a delivery center abroad, it feels like a bait and switch, whether or not anyone intended one.

Qantas isn’t alone in Australia. ABC News reported that Telstra announced a $700 million joint venture with Accenture last year, with plans to cut 209 jobs revealed earlier this year, and that Coles is reported to have up to 1,000 jobs moved offshore after partnering with the firm. Nine noted that National Australia Bank has announced plans to hire more than 1,000 staff in Vietnam and India.

What the market heard

The 6% share rise is the part I keep coming back to.

Investors don’t bid up an airline’s stock because it said “AI.” Every company says AI. They bid it up because they think costs will fall. ABC reported that Qantas last week flagged cuts to 4% of Qantas and Jetstar flights and said it could adjust fares and capacity in response to higher fuel costs. In that setting, a deal that lowers fixed corporate cost is easy to price.

That’s a real human signal, and it’s worth reading correctly. People were buying, and the reason was most likely the cost line, not a belief that AI will soon run Qantas’s finance department. If the market thought the savings depended on unproven AI, the reaction would probably have been more cautious. Labor arbitrage is proven. That’s what makes it bankable.

My take

In my opinion, Qantas’s Accenture deal is an offshoring and outsourcing deal with an AI modernization program attached, and the order of the sentences in its own statement says so. The first jobs affected won’t be replaced by software. They’ll be replaced by other people doing the same work in India, under Accenture’s management, starting in mid-2027.

The AI part is real but later, and it’s mostly about fixing the data and systems that make AI usable at all. That’s worth doing. It’s also a multi-year bet whose payoff Qantas hasn’t quantified.

The limit of my argument: I don’t know the split. Qantas said the work will be delivered “mostly” in India and “some” in Australia, and that individual outcomes haven’t been determined. If the consultation shows that a large share of the 850 roles disappear because of automation rather than relocation, the AI framing will look more accurate than I’m giving it credit for. I’d want to see that breakdown before anyone counts these as AI job losses.

Where it stands, and what to do with this

Qantas says it will consult with affected employees and their unions before individual outcomes are determined. It says it’s focused on redeployment into other roles across the group, including new roles it’s creating in Australia, plus reskilling and retraining. The union says it will fight for every job. Nobody at Mascot knows yet whether their role is one of the 850.

Hudson’s memo said the changes “are not a reflection of the contribution, commitment or performance of the teams involved.” I believe that. Work that can be written down and handed off gets handed off when the price is right, no matter how well the people doing it today do it.

My recommendation, if you’re reading about a company deal with “AI” in the headline: find the sentence that says who delivers the work on day one. If the answer is “teams in” a lower-cost country, you’re reading an offshoring story, and the AI is a promise about what happens next. And if your own job is in a back-office function that runs on documented processes, the useful move is to get closer to the parts of the work that can’t be written down: the judgment calls, the exceptions, the relationships with the people who fly and fix the planes. Those are the parts that stay.