In 2022, when we returned to the office after Melbourne's two years of lockdown, it was the silence that surprised me most. Not because the people weren't there, but because the juniors we'd hired and the ones we promoted during the pandemic were listening intently, absorbing everything around them.
As artificial intelligence reshapes the way our industry works, what I hear from colleagues and witness in my own agency is that our offices are falling silent again. Only this time it is not because the juniors aren't listening. It is because fewer of them are making it through the front door.
This is not an accident. It is a rational response to the changed economics of expertise.
A rational choice with a hidden cost
When I founded Stake 18 months ago, I made a deliberate structural choice: hire seniors and leverage AI as much as we reasonably can. Clients pay for judgment, and judgment is what senior practitioners carry. In a world where AI can produce a competent first draft in seconds, the most defensible thing a firm like ours can offer is the capacity to know what to do when the situation has no precedent and the stakes are real.
It was a sound business decision. It was also, I have come to realise, a decision that removes me from the pipeline from which I have benefited for 25 years.
Across my career, I have absorbed tacit knowledge from senior people who never had to teach it explicitly, because proximity did the work. I sat near people making hard calls. I watched them manage a crisis with no playbook, read a room wrong and recover, and hold a line under pressure when the easier path was to fold. Whatever the work was that put me in those rooms, it was never really the point. The rooms were the point. And I was allowed to be in them.
And here is what troubles me about that. The market is now actively rewarding firms for making exactly this choice. Not through malice, but through logic.
AI reduces the economic case for hiring at the base. A senior led, AI supported model is not just a positioning line, it is where the incentives point. Which means the mechanisms through which future expertise develops are being quietly dismantled, not by careless firms, but by careful ones.
The question nobody has answered
Last year, I had the opportunity to ask a chief technology officer of one of the world's most valuable technology companies how they were approaching this challenge as an employer. How do you develop judgment in people who may never have to spend a day doing the grinding work that judgment used to be built through? They didn't have an answer.
If the organisations building this technology haven't solved it, the communications industry's failure to reckon with it looks less like negligence and more like a shared blind spot. AI removes the need for the ticket. But the room still exists, and judgment still needs to transfer somehow.
The lost habit of learning by being there
The traditional apprenticeship in our profession was never formally designed. It worked because volume created proximity, and proximity created osmosis. Junior practitioners wrote the releases nobody used, compiled the media lists, and sat in on the client calls they weren't expected to contribute to.
In doing so, they absorbed, slowly and imperfectly, the thing that no brief can teach: how a senior mind moves through an ambiguous situation and arrives at a defensible position.
That system is now quietly breaking down, and nobody is building a replacement. The apprenticeship pipeline is not collapsing because anyone decided it should. It is collapsing because increasingly it doesn't make sense to pay for it anymore. And for the time being, the people who will feel that most acutely are not the firms making the rational choice. They are the future practitioners who never get to be in the room.
Making judgement visible
So, what do we owe the generation ahead of us?
Two things, I think. The first is deliberate proximity. Senior practitioners need to stop assuming osmosis will do the work and start narrating their judgement out loud. Invite the junior into the client call not to contribute but to watch. Debrief out loud after the hard conversation. Say the thing you would previously have just done.
The second is counterintuitive: use AI as the mechanism, not the replacement. Let AI generate the options. Require the junior to choose and defend and then interrogate that defence. The judgment transfer happens in the interrogation, which is closer to how good legal and medical training works than anything our industry has ever formally done. AI, used with intention, can become the forcing function for the kind of explicit, reasoned, defensible thinking we were always trying to develop but never had a reliable way to teach.
For two years our profession has been consumed by anxiety about what AI means for our own careers. That is a reasonable response to an unreasonable rate of change. The harder question, and the one we have been slower to ask, is what it means for the people coming after us.
While the future remains unclear, we can at least be deliberate about handing on what we were given.
Matthew (Matt) Thomas is Founder and Chief Catalyst at Stake The Reputation Company, a Melbourne-based consultancy working across brand, reputation, communications, and public affairs. He has advised some of Australia’s largest private companies and has worked extensively with global organisations localising their storytelling and narratives for Australian audiences. His experience spans consumer, government, health, infrastructure, technology, and corporate reputation, including advisory work at all levels of government in Australia.
Matt’s work sits at the intersection of communications, behaviour change, and institutional strategy. He is also a contributor to the The Oxford Handbook of Social Purpose, writing on reputation, legitimacy, and the growing gap between organisational messaging and operational reality.
.jpg)
-Aug-02-2026-11-32-41-5422-PM.png)
.png)