Opinion: Prioritise redesigning care around AI in health care, not adoption
AI is central to many debates across all industries, and health care is no exception. These debates and theories span from diagnostics to implementation in patient care, from supporting clinical decisions to optimising workflows.
Past conversations and trends suggest the greatest risk is expecting technology to solve problems that are fundamentally about how care is designed and delivered.
Evidence from health technology implementation shows that organisations often begin by applying new tools to existing processes. But the greatest value comes when technology is integrated with redesigned workflows, clear clinical objectives and measures of patient and service outcomes.1 This pattern has been tested across multiple sectors, such as banking, e-commerce and health care.
Transformation doesn’t start with the technology
Large-scale digital transformations commonly underperform when implementation focuses on technical delivery without sufficiently defining the behavioural, workflow and service outcomes required.
In health care, this makes clinical engagement, workflow integration and ongoing evaluation essential.1 An example of underperformance is in large enterprise systems transformation. The technology may be delivered successfully, but staff may continue using the legacy system because the organisation has not clearly defined the behavioural changes and business outcomes it wants to achieve.
Health care is approaching an inflection point in AI technology and how it’s integrated across systems. However, the stakes are much higher as failures can affect patient safety, equity and trust.
The lesson to share with the healthcare industry is simple: implementing a new system or process is not, by itself, organisational transformation. Transformation requires sustained changes in practice and demonstrable improvements in outcomes.
In health care, technology investments should ultimately be measured by their impact on patients, clinicians and services, not by whether they were delivered on time or on budget. The organisations that succeed are those that view AI as an enabler of service redesign rather than a standalone solution.
From information overload to usable insight
A major challenge in health care is the lack of usable insight at the point of care.
There is no shortage of information; however, staff shortages continue to increase pressure on clinician capacity and time dedicated to their patients.
Clinicians are expected to make decisions based on years of assessments, clinical records and patient-reported outcomes, often under significant time pressure. As information volumes grow, the time available to interpret them does not.
AI can create meaningful value by synthesising information, surfacing relevant context and highlighting meaningful changes over time.
Across health care, from mental health and chronic disease management to cancer care and hospital discharge planning, clinicians are navigating increasingly complex and fragmented information.
Administrative AI illustrates this distinction. A scribe may draft a referral quickly, but it does not necessarily resolve the consent, coordination and follow-up tasks that consume clinician time. The opportunity is therefore broader than automating one step: it is to redesign the end-to-end pathway.
The opportunity is to help clinicians make better-informed decisions, not simply automate tasks.
**************************************************
Case study: Mental health — seeing the whole patient journey
Much of health care remains organised around individual episodes of care. Yet many conditions, particularly mental health conditions, are inherently longitudinal.
Mental health support cannot focus on the snapshot of information provided in a single appointment, but must adopt a much more agile approach, considering the patient’s broader journey over time.
The value in AI lies in identifying emerging patterns across the pathway and helping clinicians recognise where additional attention may be needed.
This creates an opportunity to move from episodic and reactive models of care towards more continuous and responsive approaches.
Importantly, AI should not replace clinician decisions. Its role can be to surface relevant information, identify meaningful changes and provide a clearer picture of what is happening between appointments, while keeping responsibility for decisions with the clinician.
At a service level, the same longitudinal insights can help organisations identify population trends, target resources earlier and better understand changing demand across communities.
**************************************************
The foundations matter
Redesigning care around AI also requires getting the foundations right.
Interoperability has made information more accessible across healthcare systems, but access alone isn’t enough. Australia’s National Healthcare Interoperability Plan emphasises the need for strong frameworks around identity, standards, access and consent so that information can be shared safely and securely and used by the right people for the right purpose.2
Governance also needs to reflect the level of risk. Summarising documentation or coordinating workflows presents a different risk from clinical decision support. The Australian Government’s review of AI in health care reinforces the need for a risk-based governance approach, noting that greater potential impact on patients requires stronger oversight and risk management. In practice, this means prioritising accountability, transparency, data quality, clinical evidence, informed consent, and safeguards against human and technological error.3
AI in clinical settings should assume greater responsibility only when supported by evidence, clearly defined limits, clinician oversight and mechanisms for correction. Higher-risk functions should undergo appropriate validation and regulatory assessment before release.3
From deploying AI to proving value
The question should now focus on where AI can create meaningful value in consolidating workflows, and what evidence and safeguards are needed to support its use.
Success should be measured by whether AI delivers measurable improvements in care quality, access, clinical capacity and patient outcomes.
The organisations that create the greatest value from AI will be those that redesign care around what the technology makes possible while maintaining appropriate clinical oversight and accountability.
In systems heavily shaped by activity targets, leaders need to back the clinically right decision, even when it doesn’t increase activity.
1. Noori GH, Bibi S, Lee SW. Artificial intelligence in healthcare practice: validation, fairness, and regulatory challenges: a systematic review. Inquiry. 2026;63:00469580261478793. doi: 10.1177/00469580261478793
2. Australian Digital Health Agency. Connecting Australian Healthcare: National Healthcare Interoperability Plan 2023–2028. Published 2023. Accessed 22 October 2026. https://www.digitalhealth.gov.au/about-us/strategies-and-plans/national-healthcare-interoperability-plan
3. Australian Government Department of Health, Disability and Ageing. Safe and Responsible Artificial Intelligence in Health Care — Legislation and Regulation Review: Final Report. Published 23 July, 2025. Accessed 22 October 2026. https://www.health.gov.au/resources/publications/safe-and-responsible-artificial-intelligence-in-health-care-legislation-and-regulation-review-final-report
Why digital resilience is becoming critical in aged care
As aged care becomes more dependent on digital systems, resilience becomes inseparable from care...
How effective are ML algorithms in ICU mortality predictions?
Testing two machine learning algorithms — extra trees and gradient boosting — a team...
Back pain hybrid model aims to ease hospital bed pressure
The Back@Home trial will test a hybrid virtual and in-person care model for adults 65 years and...
