The $116 Billion AI Number Comes With a Warning Label

A big number is doing the rounds, and it deserves both your attention and your scepticism. New EY-Parthenon modelling estimates that AI could add between $95 billion and $116 billion to Australia's economy by 2036. It is the kind of figure that ends up in a board pack with a green arrow next to it. Useful, as long as it comes with the conditions attached.

A big number with conditions attached

Here is the headline, verified. AI could add between $95 billion and $116 billion to Australia's economy by 2036 and support 36,000 to 44,000 additional full-time equivalent jobs, according to EY-Parthenon modelling. (Source: EY, 'AI productivity gains could deliver up to $116bn boost to Australia's economy', August 2026). The uplift would be underpinned by a 2.0 per cent to 2.4 per cent increase in multifactor productivity and $31 billion to $38 billion in additional investment. (Source: EY, 'AI productivity gains could deliver up to $116bn boost to Australia's economy', August 2026).

The reason a productivity number lands so hard right now is that the recent track record has been poor. Labour productivity growth had averaged just 0.3 per cent a year in the past decade (Source: AAP via Michael West Media, 'Report reveals extent of AI boost to Australian economy', August 2026), and multifactor productivity, which measures how well labour and capital combine to produce outputs, declined 0.5% over the 12 months to June 2025, below the 20-year average of 0.4% growth per year, according to the Productivity Commission's Annual Productivity Bulletin 2026. (Source: HCA Mag, 'AI could create up to 44,000 jobs to Australian economy, says EY', August 2026).

So the opportunity is real. The catch is in one word EY uses repeatedly: could. The modelling assumes businesses turn the technology into investment, capacity and output. Left as a set-and-forget tool, it adds very little. EY's own economist made the point plainly: the size of the benefit depends on how effectively firms reskill their people (Source: AAP via Michael West Media, 'Report reveals extent of AI boost to Australian economy', August 2026).

The gap between the macro number and your P&L

$116 billion is a whole-of-economy figure. No single business banks it, and the gains are not spread evenly. AI adoption is expected to reshape labour demand across sectors, with construction recording the largest jobs increase while agriculture and mining are expected to require fewer workers under the modelled scenarios. (Source: EY, 'AI productivity gains could deliver up to $116bn boost to Australia's economy', August 2026). A macro tailwind is not a business plan.

The more useful question for a board is where your business sits between the potential and the captured. On current evidence, most sit closer to the start line than the finish. Only 14% of organisations have scaled AI across their business, and AI use among workers hasn't budged: 46% report weekly usage in 2025, virtually flat from 2024. (Source: Asana, '2025 State of AI at Work: Australia'). Broader still, around 12 per cent of Australian businesses reported the use of AI in their workplace in 2024-25 (Source: Australian Bureau of Statistics, 'Business adoption of Artificial Intelligence accelerates in 2024-25', June 2026).

And the businesses that have adopted it are not automatically ahead. The pattern of high adoption and low return is now well documented. MIT's Project NANDA released a July 2025 report, The GenAI Divide: State of AI in Business 2025, finding that despite $30 to $40 billion in enterprise investment, 95% of generative AI projects yield no measurable business return. (Source: MIT NANDA, 'The GenAI Divide: State of AI in Business 2025', July 2025). That is an execution result, not a technology fault. The tools work. The wiring around them often does not.

What boards should do differently

Being on the right side of that gap in five years, rather than reading about it in someone else's case study, comes down to a handful of unglamorous decisions.

• Set the P&L number first. Before funding a build, name the outcome it is meant to move and how you will measure it. A pilot without a defined result is a hobby with a budget.
• Fund the plumbing, not the demo. Most value gets stuck in the boring layer: data quality, process redesign and integration into how work actually happens. That is where the return lives, and where budgets tend to run thin.
• Start narrow, then scale. The strongest efforts start with one high-value workflow, prove measurable impact, and expand from there. Breadth without depth is how you end up with 46 per cent weekly usage and nothing on the earnings line.
• Treat skills as capital spend. The reskilling point is not soft. 44 per cent of Australian CEOs rank large-scale reskilling as a top workforce priority over the next three years, while only 20 per cent prioritise reducing hiring. (Source: EY, 'The economic upside of Artificial Intelligence for Australia', August 2026). Redesigning the work is the job, not an afterthought.

The $116 billion is what the country could collect if enough businesses do this properly. Your share depends entirely on the choices in your building, not the ones in the modelling.

The right next steps for you depend on where you are right now. Our AI Maturity Assessment is an easy place to start. After that, let's have a chat about your next steps.