AI agents set to wipe out middle managers, says Quanton chief

Quanton managing director Garry Green says AI agents are set to replace layers of middle management in Australian businesses, leaving smaller human teams to oversee far more autonomous systems.

He argues that agentic AI marks a break from earlier workplace automation because the software can run multi-step processes with limited supervision. In his view, that challenges the long-standing model in which companies grow by hiring more staff and promoting workers through a management hierarchy.

"Unlike earlier AI technologies that simply assisted human workers, agentic AI can independently manage entire workflows from start to finish," Green said.

He says these systems could take on work such as contract negotiations, supply chain optimisation and customer dispute resolution-tasks that have often sat with trained professionals or managers.

Middle Layer

At the centre of Green's argument is the idea that the traditional corporate pyramid is losing its middle tier. As AI systems improve, he says, both entry-level analytical work and the coordination work usually handled by middle managers are becoming more exposed.

"Businesses have always been organised as hierarchical pyramids-graduates at the bottom doing analytical work, slowly moving up through middle management ranks as they gain experience," Green said. "That entire structure is collapsing. AI agents can now handle both the entry-level analytical work and the middle management orchestration that used to require years of human experience. You simply don't need that middle management layer anymore."

Green says the economics are straightforward: AI systems can operate continuously and scale without the salary costs attached to human recruitment. He describes a future in which businesses increase output while reducing headcount.

"A business that previously needed 50 people to handle a certain workload might need five humans orchestrating 100 AI agents. Same output, dramatically different cost structure."

The claim reflects a wider debate in business and policy circles about whether AI will eliminate roles outright or simply reduce hiring over time. Green says companies may increasingly rely on attrition rather than redundancies, leaving departing workers unreplaced while software takes on their tasks.

Legal Example

For example, Green points to a contract review conducted within Quanton. He says an AI agent reviewed a Master Service Agreement against standard terms and prepared amendments in about an hour, a process that would usually take days of exchanges with lawyers.

"If an AI agent can deliver results better, faster, and with more precision than an internal legal team, that team becomes a difficult commodity to justify," Green said.

He links that example to a broader shift in professional services, where early-career legal and analytical work has traditionally served as a training ground for future senior staff. If those entry points are increasingly automated, the path into senior white-collar jobs could narrow.

New Skills

Green says the shift will also change the type of worker companies need. He describes the rise of what he calls the "M-shaped" worker: someone with broad knowledge across several domains and the ability to go deeper when required.

"Going forward, your breadth of knowledge has got to increase," Green said. "It's no good to have a siloed understanding of things, because you won't be 'managing a particular function' in the business, you'll be solving problems and directing AI agents to achieve business goals."

"Work will shift from 'doing the job' to 'orchestrating an outcome'."

That would mark a departure from the "T-shaped" model often favoured in corporate hiring, where workers build deep expertise in one area and a lighter understanding of adjacent fields. Green says supervising AI systems effectively will require a broader grasp of end-to-end business processes.

Political Pressure

Green says the labour-market effects are already becoming politically sensitive, particularly as white-collar occupations come under pressure. He argues that public debate has yet to catch up with the way companies may adopt AI without formally declaring job losses.

"They're talking about legislation requiring companies to report where AI replaces a job," Green said, "but that's just naïve. Smart businesses will simply stop hiring. As people leave, those roles won't be replaced and they'll be filled by AI agents. Businesses will grow through automation rather than recruitment."

"There's nothing new about that. It's restructuring through attrition, and it's already happening."

He says the social effects could be sharper because the workers affected are likely to include lawyers, analysts and managers rather than the manufacturing or retail staff more commonly associated with automation debates.

"Unemployment triggers social unrest-that's historically proven," Green said, "but what's different this time is that it's not factory workers or retail staff who will be affected. It's lawyers, financial analysts, middle managers. White-collar professionals who thought they were immune to automation are now in the firing line."

Green argues that business leaders should redesign organisations for flatter structures rather than simply adding AI tools to existing hierarchies.

"This is truly exponential technology," he said. "AI models are growing 10x in capability every few months. That's a massive acceleration. People underestimate exponential change-one minute they're doing things the way they've always been done, then suddenly: boom, everything changes."