New research from Equal Experts APAC found structural and data maturity gaps are slowing AI adoption in Australian healthcare. The study drew on conversations with more than 100 digital health leaders across the sector.
The findings identify organisational design as the biggest barrier to moving artificial intelligence beyond pilot programs. In the survey, 65% of respondents rated their organisation's operating model at 1 or 2 out of 5, the lowest score across the areas assessed.
Participants said traditional silos and decision-making processes have not adapted to the speed and iterative approach required for AI deployment. Without changes to those structures, the research suggests many projects will remain at the experimental stage rather than delivering measurable clinical or operational results.
Data barriers
Data quality and infrastructure also emerged as major constraints. The study found 40% of respondents rated their organisation's data management maturity at 1 or 2 out of 5, while 38% gave equally low scores to their tools, platforms and technologies.
Legacy systems were identified as a recurring problem. Outdated technology often prevents AI models from accessing the real-time, high-quality datasets needed for decision support, creating integration failures and limiting broader deployment.
The report was based on six months of interviews, survey responses, executive roundtables and industry panels. Participants came from hospitals, primary care, aged care and health networks, giving the study a broad view of how the sector is approaching AI adoption.
Governance concerns
Governance and oversight were another weak point. One-third of respondents rated their organisation's ethics, governance and monitoring at 1 or 2 out of 5, while 50% said their AI lifecycle management remained immature.
The research also highlighted concern about so-called shadow AI, in which staff and clinicians deploy AI tools without formal organisational oversight. It identified this as partly the result of gaps in operating models and governance, leaving institutions without clear controls as use of generative and analytical AI expands.
Andy Canning, Chief Technology Officer and Managing Director, Equal Experts APAC, said organisations needed to focus less on experimentation alone and more on the systems that support long-term adoption.
"The AI landscape in healthcare already looks radically different than it did 18 months ago, and in another 18 months, it will be unrecognisable. However, leaping into AI without the right foundations across data, governance, and clinical culture is a high-risk strategy, leading to stalled projects and wasted investment. The organisations building solid foundations today will be the ones positioned to maximise the value of AI in the future," Canning said.
The results suggest interest in AI is not the main issue for Australian healthcare providers. Instead, the sector appears to be facing the harder task of reshaping internal processes, updating technology estates and putting governance in place before tools can be adopted at scale.
That challenge is particularly acute in healthcare, where fragmented data environments, strict compliance requirements and complex clinical workflows can slow adoption more than in other industries. The report indicates many organisations are still at an early stage of readiness despite growing pressure to use AI in both clinical and operational settings.
Ken Gallacher, Healthcare Executive Advisor, Equal Experts APAC, said the research showed a divide between providers treating AI as a narrow technology initiative and those viewing it as broader organisational change.
"We've been privileged to speak to healthcare leaders at all stages of AI maturity during our research. We are seeing a clear divide between organisations that treat AI as a standalone IT project and those that treat it as a fundamental shift in how care is delivered. To move past single pilots, healthcare providers must break down traditional silos and embed AI governance across the entire people-and-process lifecycle," Gallacher said.