Australia targets up to AUD $73.3 billion boost for AI by 2028

Australia is set to expand its focus on artificial intelligence, with an anticipated up to AUD $73.3 billion in private investment in data centres by 2028 under the National AI Capability Plan.

Federal Industry and Innovation Minister Tim Ayres has outlined ambitions for the plan to be "broader and more expansive" than previously outlined, reflecting a national commitment to increasing AI capability and adoption across sectors.

AI adoption and uneven returns

Recent research from the Massachusetts Institute of Technology (MIT) indicates that despite rapid uptake of artificial intelligence solutions by Australian businesses, many are not achieving the return on investment they anticipated. The reason, as highlighted by both MIT and Wharton School research, lies in the complexity and quality of data available to businesses.

Large language models (LLMs), such as ChatGPT and Claude, function well when drawing on structured data, facilitating commonplace business tasks. However, the use of AI in more complex business processes is often hampered by fragmented data systems, particularly in industries grounded in physical production and outdated IT infrastructure.

"Industries more anchored in the world of real things face fragmented data, legacy systems, and physical-world friction," the Wharton study found. "It's not that AI doesn't work for manufacturers. It's that their data isn't ready for it."

The result, as reflected in the Wharton report, is a disparity in positive returns on AI investment: while banking and technology sectors report positive returns on investment at rates between 83% and 88%, manufacturers achieve 75% and retailers are at just 54%.

Manufacturing's data dilemma

This divide is attributed to the inherent challenges manufacturing companies face in digital transformation. With multiple, often siloed, management platforms for customer relationships (CRM), enterprise resources (ERP), human resources (HR), and finance, integrating data is particularly arduous in these sectors.

Professor Cori Stewart, Chief Executive Officer and Founder of ARM Hub, commented on this issue among Australian manufacturers: "The technology works fine. But if your data lives in five different places with three different versions of the truth, no AI tool will magically fix that."

Government support and practical steps

Government strategy includes significant investment in the AI Adopt Programme, delivering AUD $17 million to establish four AI Adopt Centres which provide hands-on support and tools for businesses implementing AI technology. As Minister Ayres stated, the National AI Capability Plan is "built on three core principles," with the first being "purposefully capturing the opportunities of AI."

"Businesses who have engaged with these centres are already finding ways to improve the world of work and service provision-reducing intensity for some workers and improving clinical accuracy for others," Minister Ayres said.

The plan also prioritises building "trust and confidence" in AI technology and ensuring that its benefits are distributed equitably.

Incremental fixes in data management

Case studies from the manufacturing sector underscore the potential for incremental gains from consolidating and organising business data. Brisbane-based UAP reduced its project forecasting time from three days to a few hours by working with ARM Hub's AI Adopt Centre to streamline its scheduling processes. The Air Conditioning and Mechanical Contractors' Association has deployed conversational AI systems that provide straightforward compliance information and automate service reporting, all without forming in-house data science teams.

These companies benefitted from ARM Hub's Data and AI-as-a-Service, assisting with centralising information and establishing interoperability between previously disconnected systems.

Professor Stewart discussed this approach when addressing delegates at the Gladstone Engineering Alliance Major Industry, Energy & Manufacturing Conference: "Businesses think AI adoption means a massive transformation project. Our approach breaks it into manageable steps. Pick one problem that matters to your business. Solve it. Then build from there."

Her suggested framework includes beginning with quality, organised data and setting up projects with measurable returns on investment, maintaining workforce involvement, thorough testing, and continual monitoring. This model addresses the key reasons why manufacturers and other sectors with less digital maturity struggle to achieve equivalent returns from AI utilisation.

Sector-specific challenges remain

The disparity identified in the Wharton research suggests that digital-native industries such as banking and technology are better positioned to monetise AI as their data is already structured and accessible. Manufacturing, however, is likely to require ongoing support to overcome issues of fragmented and inconsistent data structures before AI solutions can deliver full value.

For businesses struggling to centralise their data or determine initial AI priorities, ARM Hub continues to provide hands-on guidance and has now worked with more than 300 Australian firms seeking to build capability in this area.