The AI innovators helping Australian companies get ahead

The AI innovators helping Australian companies get ahead

Adelaide University researchers are making major strides in AI science, developing practical, tailored advancements for industry partners – and reaching new heights of efficiency.

AI is transforming industries the world over. The smartest companies are using it to automate tedious, repetitive tasks at work, freeing up more time for strategy, creativity and innovation. The real competitive impact comes from bespoke AI advancements that speed up systems – driving productivity and efficiency for business owners and customers alike. 

This is the focus of research at the world-leading Australian Institute for Machine Learning (AIML), where multiple AI initiatives are making waves across the commercial sector. The first of these is a flagship partnership between AIML and Commonwealth Bank of Australia (CommBank) called the CommBank Centre for Foundational AI Research. ‘Foundational AI’ explores the mathematics and algorithms underpinning AI and uses those insights to make it more efficient. 

The partnership is already delivering incredible results, with Adelaide University researchers developing an innovative solution for a major CommBank challenge. 

Banks rely on AI to identify fraud, assess risks, and understand customers. But with sensitive financial data and millions of transactions to manage and protect, they can’t use standard open AI models for this. They need fast, accurate systems that keep information secure.  

At AIML, Professor Simon Lucey and Dr Hemanth Saratchandran tackled this challenge by rethinking the maths at the core of the model – rather than scaling up infrastructure. 

“We shifted model updates into a richer, non-linear mathematical space to increase power without increasing compute, memory requirements, or infrastructure costs,” Professor Lucey says. “By fine-tuning the models, they became smarter and more adaptable, without becoming more expensive to run.”

The solution required no new hardware, no additional expenditure, and no increase in energy use. 

The results were significant. Computing requirements dropped by 50%, directly reducing energy consumption and infrastructure load. Training was faster, meaning shorter development cycles and quicker iteration. The approach improved predictive performance on the bank’s financial data, supporting more reliable customer insights. 

“AI adoption is no longer exclusive to tech companies with deep pockets and specialised data science teams.”

Jonathon Read
Australian Institute for Machine Learning

Rather than putting resources into existing tools, CommBank and Adelaide University investigated the underlying science of AI – and reaped the rewards. The more accurate AI directly translated into better protection for customers, stronger risk management, more sustainable outcomes, and a more agile organisation.

“This collaboration demonstrates how strategic investment in foundational AI research can deliver rapid, measurable returns,” Professor Lucey says.

In 2025, AIML also officially launched the Industrial AI Grant Program, created to help businesses explore AI adoption in their operations. While many companies know they would benefit from AI automation, they often lack the expertise to implement it or the certainty that it’s worth the risk. 

The Industrial AI Grant Program offers two kinds of support to help with this. The first is the AI Roadmap Generator, a tool that provides a starting point for business to consider how AI and machine learning might benefit them.  

“Close to 100 businesses have used the roadmap to start adopting AI in their processes,” says AIML Engineering Manager Jonathon Read, who assists Principal Engineer Stefan Podgorski in leading their team of a dozen engineers. 

“The process helps them understand where AI can add value and supports them to build a business case for its use.”

The ML Innovate stream, a second offering of the Grant Program, focuses more on hands-on problem solving. Participating businesses gain direct access to the AIML engineering team, led by Podgorski, to help ideate and test possible AI solutions in their industry. By trialling before investing, businesses can reduce the risk of AI adoption, all while upskilling their staff. 

There have already been many success stories. For example, AIML engineers partnered with agtech company Cropify to develop computer vision-based automation for grain and pulse quality assessment, significantly accelerating the process – from a 24-minute manual process to 90 seconds – and improving consistency. They also worked with Digital Constructors to create AI that automatically detects infrastructure issues on train lines – technology with the potential to save millions in maintenance costs. Corporate Threat Intelligence company Sention leveraged AIML-developed tools to enable automated classification and ranking of corporate risks, conditioned on each organisation’s specific threat profile. These insights allow organisations to respond to emerging threats more quickly and effectively. 

On the horizon

AIML’s AI breakthrough will continue to scale across CommBank, and its implications extend far beyond banking. Any organisation that holds rich, sensitive, domain-specific data – and needs AI that truly understands it – stands to benefit from the advanced fine-tuning method. In hospitals, AI trained on patient records could improve diagnoses, predict deterioration and streamline care. For insurance companies, more accurate risk models would mean fairer premiums and faster claims. Telecommunications providers could benefit from real-time pattern recognition, and all government agencies require secure environments to hold sensitive national datasets. 

The continued aim of the Industrial AI Grant Program, meanwhile, is to build a sustainable ecosystem of AI-capable businesses – and a collaboration model that’s easy to replicate. The program is strengthening capacity for the long term, supporting the next generation of AI leaders and contributing to the advancement of industrial AI both nationally and globally. 

“AI adoption is no longer exclusive to tech companies with deep pockets and specialised data science teams,” says Read. “Through initiatives like ours, all levels of business can access world-class machine learning expertise.”  

This democratisation is where the real impact lies: in making AI accessible and practical for businesses across all sectors. As adoption becomes easier and processes more efficient, industries around the world can streamline operations, strengthen data security and innovate at a pace like never before.