Journal logo

The Process Nobody Wrote Down: What the Qantas Outsourcing Deal Gets Right and Risks

Using a partner to move faster on AI is sensible, as long as the knowledge behind the work stays inside the company.

By Ashkan RajaeePublished a day ago • 4 min read
The Process Nobody Wrote Down: What the Qantas Outsourcing Deal Gets Right and Risks
Photo by Hanson Lu on Unsplash

Note: This story was written with AI assistance and edited by me.

Picture a company that hands its whole billing process to an outside team. The handover takes two weeks. Untangling it takes most of a year. Nothing about the new team is weak. The problem is that the one person who knew why invoices for a certain group of customers were held back every month has already left, and that rule was never written down. The new team follows the documented steps perfectly and sends out invoices that should never have gone. Anyone who has worked in software services long enough has seen some version of that story.

I thought about that pattern when I read the recent news from Australia.

What happened

On November 10, 2026, Qantas announced a multi-year partnership with Accenture to modernise its back-end systems, processes and data. According to the airline, it runs on more than 1,000 legacy applications that were never designed to work together. Around 850 roles are expected to be affected over the coming years, starting in corporate functions such as finance, procurement, people and marketing, with the first changes expected by mid-2027. Qantas said some of those services will be delivered mostly by Accenture teams in India and some in Australia, and later reshaped with automation and AI.

Chief executive Vanessa Hudson said the company could not hire fast enough on its own, adding that a partner gives it access to the data and AI skills "that every industry in the world is competing for." The Australian Services Union criticised the move, and Qantas said it would consult staff and unions and look for redeployment where possible.

That is the news as reported. What follows is my view.

Why the reasoning makes sense

I do not think the decision to use a partner is the mistake here. Skills in data engineering and production AI are genuinely scarce, and an airline is not built to recruit hundreds of those people in a year. Bringing in outside capacity is a normal, often sensible way to get modern systems built while the business keeps running.

Location is also not the real issue for me. I have spent most of my career working with distributed teams, and good engineers deliver across time zones every day. A process run well in Bengaluru is better than the same process run badly in Sydney.

The part that worries me

What worries me is timing. In most transformations like this, the process moves at the same moment the people who understood it leave. Old workflows are full of small judgment calls: the supplier who always gets paid early, the approval that exists because of a problem from ten years ago, the manual check that catches one error a quarter. None of it shows up in a process map.

When that knowledge walks out the door, the receiving team inherits the steps without the reasons. Then automation gets layered on top, and the hidden mistakes simply happen faster and at greater scale.

There is a second risk. If the redesigned process is meant to return later as an automated function, someone inside the company has to be able to judge whether the new version is actually better. If everyone who knew the old version is gone, nobody is left to ask the hard questions.

How I would approach a move like this

If I were leading a change of this size, I would start with a few simple rules.

First, keep one named owner inside the company for every process being moved. That person does not do the daily work anymore, but they approve the redesign and answer for the result.

Second, spend the first weeks collecting exceptions. Sit with the people doing the job and record every workaround, manual fix and unusual rule before anything is transferred.

Third, judge success by outcomes. Look at error rates, cycle times and customer complaints, not at how many roles have been moved.

Fourth, pair internal staff with the outside specialists for a period, so learning goes in both directions.

Fifth, tell people what is happening as early as you honestly can. Staff who feel blindsided tend to stop sharing what they know, and their knowledge is the most valuable thing in the room.

The other side of the argument

It is fair to say Qantas is not shrinking overall. The airline says it has created more than 4,400 operational roles since 2023 and still calls itself a net employer. It is also facing high fuel costs, so some of this is plain cost pressure rather than a grand AI strategy. And a partner with deep benches can move far faster than any internal hiring plan. I am not arguing against the deal. I am arguing about the order in which the work gets done.

What I have learned

Earlier in my career, as a partner, I helped grow a software consulting firm from about $6M to $60M in under three years. Across our client work, one pattern held. The engagements that went well had a strong owner on the client side, someone who understood the business deeply enough to push back. The ones that struggled had handed over both the work and the thinking.

A simple place to start

If your company is considering outsourcing part of its operations to speed up AI adoption, pick one process this week and ask the people who run it to list every exception they handle in a typical month. If that list surprises you, you have found the work that needs to happen before anything moves.

careerbusiness

About the Creator

Ashkan Rajaee

Enjoyed the story? Support the Creator.

Subscribe for free to receive all their stories in your feed.

Subscribe For Free

Reader insights

Comments

There are no comments for this story

Be the first to respond and start the conversation.

Sign in to comment
    Written by Ashkan Rajaee