2 days ago
Agentic AI is Australia's unrealised productivity lever
As ongoing Productivity Commission reports show, Australia's flat productivity is a persistent national challenge, so much so that the Labour government is hosting a three day industry roundtable focused on restarting our economic engine. If productivity is corrected, the Productivity Commission estimated that it would leave the average Australian worker $14,000 better off by 2035.
However, simply working harder isn't the answer. The latest ABS Labour Force data highlights that Australians are working ever more hours. We collectively clocked 2 billion hours in July 2025, which is a 2.1% annual increase. Yet, our national productivity stagnates. Instead, we should be working smarter.
As the Economic Reform Roundtable convenes in parliament, it's positive to see the Treasurer put more emphasis on AI as a key lever to solving Australia's productivity deficit - estimating AI could add more than $116 billion to the economy over the next decade. But the national conversation is still catching up to reality.
While policy debates dwell on generative AI, the frontier has already evolved to agentic AI: systems that plan, action, and collaborate with humans across workflows. The risk isn't moving too fast; it's moving too slow and falling behind the global status quo.
With Microsoft's Work Trend Index 2025 showing that 81% of business leaders globally expect AI agents to be part of their strategy in the next 12–18 months, it's crucial for Australian industry and government to recognise agentic AI as a critical asset in not only course-correcting our productivity, but moving us ahead economically too.
However, like any transformational technology, the fundamental challenge is skills; agentic AI requires new skills and a rethink to how we approach work. Fortunately, there's already a lot the white collar workforce can learn from how the software development world has reinvented itself with agentic AI.
Software development's blueprint for agentic AI success
The software development industry provides a powerful blueprint for agentic AI adoption, particularly for a market like Australia that suffers a perpetual tech skills shortage which puts consistent pressure on developers to deliver more and more.
AI coding agents are opening up doors for human developers to have their own agent-driven team, all working in parallel to amplify their work. Technical teams are now able to assign tasks that would typically detract from deeper, more complex work, allowing developers to focus on high-value coding tasks that have the most impact.
We've seen a profound impact on the use of GitHub Copilot's agentic capabilities, with many now using the tool in their day-to-day work. Australian industry leaders like Commonwealth Bank, EY, and MYOBhave embraced AI-assisted software development, leading to surging productivity gains.
More broadly, between 60-75% of developers report feeling more fulfilled and less frustrated when using GitHub Copilot - a metric businesses shouldn't take for granted when striving to maximise productivity.
By taking on repetitive, time-consuming tasks asynchronously, AI agents free developers to focus on what matters: solving problems, delivering value, and building products at scale. What once took days now takes hours, turning release cycles from marathons into sprints, offering up a surefire way to reshape the productivity puzzle.
What can businesses learn from this? Employees who are empowered with AI,integrating it deeply into their workflow, emerge with greater ambition, sharper technical fluency, and higher job satisfaction. The evidence shows that developers aren't writing less code; they're orchestrating more complex, system‑level work, and this should provide an exciting example for businesses to follow. It will take a lot of 'unlearning' at the senior manager level (and a lot of trust in employees) but a 'reimagining of work' across a business' workforce will open creative and innovative doors that have probably not been conceived of yet.
For this opportunity to be harnessed across Australia's economy however, we need to instill these skills in future generations; 'managing AI' should be a fundamental skill every citizen learns.
AI education is the key to empowering an innovative society
We can't expect graduates to thrive in an AI-driven economy if they haven't learned with AI. It's the equivalent of trying to prepare the future workforce to work with computers, apps, and other modern digital means, and then handing them a pencil.
Australia has taken an important first step with the Australian Framework for Generative AI in Schools, recently endorsed by education ministers in June 2025. But a framework centered on GenAI alone is already being outpaced by the next frontier with agentic AI.
Students should be given the opportunity to learn about AI as a dedicated subject in school curriculums. They should be taught to brief, supervise, and evaluate autonomous systems: the very skills that will be critical to their future jobs on day one. Never before has technology been so ubiquitous and instrumental in the future of the economy - leaving it out of early learning would not only do a disservice to students, it's also not conducive to a competitive economy.
This isn't a theoretical risk either. Australia is already struggling to attract and retain AI talent. We lean heavily on skilled migration for advanced roles; we punch above our weight in research output but capture only a sliver of AI patents, a clear sign of a commercialisation gap. Too many talented graduates, including those who've helped build world-class systems, are taking their skills offshore. The fix is straightforward: create clear, supported pathways from classroom to career so students can learn, build, live, and work in Australia.
As we await tangible policy following the Economic Reform Roundtable, it's imperative that both government and industry understand one thing: productivity is intrinsically linked to economic growth - and AI sits at the heart of this.
If Australia wants to compete, the choice is simple: reinforce agentic AI as a productivity lever and reinvest in the human capability that unlocks it. While the Roundtable can set the direction, what remains is the will to link policy with practice so investment and education moves together toward reversing the trend of declining national productivity, and making our economy world-leading with AI.
