Australia's AI moment: the opportunity in front of us

Soon after ChatGPT’s launch, Canva co-founder Cliff Obrecht gave the AI tool an unorthodox test: plan a global camel delivery service, including logistics, routing and turnaround times.

The model nailed it. It laid out a complete operational plan, even though the challenge was so abstract that Obrecht had expected it would break the technology.

At that moment, Obrecht knew what the results meant.

"If we weren't going to disrupt ourselves," he told the audience at CommBank's Accelerate AI event in Sydney on Tuesday, "we were going to be disrupted."

The event brought together 800 business leaders to examine Australia's pace of AI adoption, and the picture was sobering. By some measures, European companies are adopting AI at twice the rate of Australian ones, and US companies at three times the rate.

The message from the day's panels was consistent.

"If you don't adopt AI, almost certainly you will lose market share,” said Seek founder and tech investor Paul Bassat. “And ultimately, you will go out of business."

Lorikeet co-founder Jamie Hall said about 60 per cent of his team’s sales were in the US, compared with 25 per cent in Europe and 15 per cent in Australia.

Distance travelled: 1 per cent

“This is unquestionably the most significant industrial transformation in modern history,” said Craig Scroggie, CEO and managing director of NEXTDC, a data centre operator.

“We are now entering the fourth industrial revolution, the production of intelligence.”

The top four hyperscalers, the US tech giants building the large language models that power generative AI, have spent more than $US700 billion between them so far this year on data centres and other digital infrastructure. The total bill for 2026 is expected to top $US1 trillion.

The level of capital pouring into AI companies and supporting infrastructure is estimated to be 2 per cent of global GDP, comparable to the investment in building railroads across the US at 6 per cent or laying telecommunications networks for the internet at 1 per cent.

For Australia, the news is mostly good. The country has a stable regulatory environment, abundant renewable energy and a skilled workforce. As grid-sized batteries and solar push down the price of electricity, Australia is an attractive location for building the many data centres required to support the intense computational requirements of an AI-operated economy.

The infrastructure investment alone could generate 100,000 jobs in the build phase, according to McKinsey senior partner Angus Dawson.

The question is whether Australia moves fast enough to claim its share.

So why does it not feel like we are living through an industrial revolution?

Despite these astounding flows, the impact of AI on daily life, jobs and the structure of society to date has been minimal.

"If the ultimate measure is impact," Bassat said, "we are less than one per cent of the way in".

Bunnings’ DIY Buddy

AI is already turning up in our lives, such as the weekend trip to Bunnings.

Earlier this year the hardware retailer launched Buddy, an AI shopping assistant, making it one of the first Australian retailers to deploy the capability at scale. Last week alone, 25,000 customers used it to get help with DIY projects, from painting a room and mowing a lawn to renovating a kitchen.

Inside stores, Bunnings deployed a staff chatbot across 400 sites and 50,000 employees. It has now answered more than four million questions, saving staff an average of five minutes a day.

Bunnings chief customer officer Rachel McVitty said the decision to move came partly from watching staff use their own personal ChatGPT tools on the shop floor. The appetite was already there.

"The sooner we got on that learning curve as an organisation," she said, "the sooner we could move up it."

The principles Bunnings applied are straightforward and transferable to any business: customers always know they are talking to AI; the tool draws only on verified content; and it never blocks a path to a real person.

The company of one

While Australia’s largest companies are launching digital employees, the greatest opportunities could be among the country’s smallest businesses.

Bassat described an emerging model: an individual, perhaps recently made redundant, setting up a business of one and running as many as a thousand AI agents doing work that previously required an entire team.

It has never been cheaper or easier to launch a global company: website, financial services, software stack and now capable AI tools, all accessible without significant capital.

The Australian government needs to support these ambitions because on a net basis, almost all new jobs in Australia are created by companies five years old or younger. Established companies, as a group, shed jobs over time.

AI will accelerate that trend by driving efficiency across large organisations, making the pipeline of new businesses more important now than ever.

Unfortunately, Australia has been travelling in the opposite direction. The rate of new businesses being started in Australia has dropped around 20 per cent over the past decade, said Obrecht. Sole trader numbers are also down.

"We need the risk takers," he said. "It's the risk takers that build the businesses that end up employing people."

You can't hire your way to an AI strategy

In organisations of any size, the greatest mistake is to think you can outsource AI adoption to a “transformation team” or the IT department.

All company-wide transitions need to start from the top. Leaders cannot delegate their way through it.

Using AI tools deeply and regularly, not nominally, is the only way to understand what they can and cannot do.

"You learn by doing," Obrecht said. "Go head first, try to solve real problems, run into dead ends."

McKinsey’s Dawson offered a useful frame for any business owner trying to work out where to start. Most organisations, he said, are focusing all their energy on the third of three essential questions: how do we mobilise?

The first two, do we understand what AI is doing to our industry, and have we decided what to do about it, are often skipped.

But this is a mistake because the answers show you where to focus your efforts in adopting AI. In a country already moving slower than its major trading partners, skipping those sorts of questions is a luxury Australia cannot afford.