I just got back from HIMSS 2026, and honestly, the energy in Las Vegas was something else. Thousands of health IT professionals packed into the Venetian, but no matter where you turned, keynotes, panels, the massive expo floor, or just grabbing a coffee in the hallways, one topic kept coming up again and again: agentic AI.
If the last couple of years were all about generative AI hype (you know, tools that spit out text or images on demand), then 2026 feels like the real turning point. We’re now seeing systems that don’t just answer questions; they can reason, plan, adapt, and actually carry out entire workflows on their own, while staying inside proper guardrails.
One keynote speaker summed it up perfectly: we’ve moved from predictive to generative, and now it’s the era of agentic AI, systems that don’t just suggest things, they do them.
For Australian healthcare IT professionals battling staff shortages, systems that don’t always talk to each other and mountains of admin, this shift isn’t just exciting. It could genuinely change how care is delivered across the country.
So what is agentic AI, really?
It’s a big step up from the chatbots we’ve been playing with. These agents can take a complex goal, like sorting a referral or chasing a prior authorisation, break it down into steps, pull information from different systems in real time, make rule-based decisions, carry out actions (submitting forms, updating records, sending notifications), and even loop back to refine things if needed. They only flag a human when something falls outside their scope.
The great part in healthcare? Clinicians stay firmly in control of the clinical decisions. The AI simply clears away all the repetitive friction that keeps them away from patients.
What I saw on the floor
The major platforms showcased new tools that let any health service build and manage its own agents directly within their core systems. These agents come loaded with local policies and knowledge, so they actually fit the way each organisation works.
The early results people were sharing were impressive: discharge summaries done much faster, prior authorisations slashed from days to minutes, patient reminders and scheduling running on autopilot. One survey that kept popping up showed 77% of healthcare organisations are already
investing in agentic AI, with most planning to go live within the next year. And the constant message? Strong interoperability is the make-or-break factor.
I also heard some really encouraging real-world stories: hospitals using agents to transform patient contact centres, speed up revenue cycle work, and make care coordination feel almost seamless.
What this could mean for Australia
What really struck me is how well-placed Australia is for this technology. Thanks to My Health Record, the National Digital Health Strategy and widespread FHIR adoption, the country already has a lot of the clean, connected data foundation that agentic AI needs to actually work.
The implications feel pretty significant:
● Workforce relief and less burnout. Australian clinicians are spending up to half their time on paperwork: referrals, medication reconciliations, NDIS approvals, aged-care reporting, and rostering. Agents could take over a lot of that automatically. Early overseas numbers suggest 10–20% savings in admin time, which could mean more face-to-face care and a real dent in burnout, especially in regional and rural areas where staffing is always tight.
● Better patient access and experience. Imagine 24/7 personalised outreach in multiple languages, proactive care-gap closing, and smooth handovers between hospital, GP, specialist, and aged-care settings. In a country as geographically spread out as Australia, this could make care faster and fairer, particularly for remote, Indigenous, and older communities.
● Financial and operational resilience. Revenue-cycle agents can chase claims, cut denials and optimise billing in real time. That matters a lot when both public health services and private providers are working under tight budgets.
● Clinical safety and quality. When done well, agents enhance decision support by pulling data from multiple sources. But the speakers were clear: they amplify whatever processes you already have, so organisations need to fix any broken workflows first.
Of course, it’s not all smooth sailing. The conference hammered home a few realities Australian leaders can’t ignore:
Governance and human oversight are non-negotiable. Strong audit trails, clear escalation paths and clinician sign-off will be essential to stay aligned with Australian Privacy Principles, TGA expectations and clinical safety standards.
Data quality and interoperability remain foundational. These agents are only as good as the information they can access.
Validation and safety checks have to keep pace. The tech is moving fast, so local testing, bias audits and clear accountability are essential.
And change management will be huge. This isn’t about replacing people, it’s about redesigning roles. Organisations that invest in training and bring their teams along will win; those that don’t risk frustration and tool sprawl.
Looking ahead
The big takeaway I brought home from Las Vegas is that agentic AI is quickly moving from “interesting pilot” to “core part of how we run healthcare.” The organisations seeing the fastest wins are starting small with high-volume areas like patient communications and revenue cycle, while building solid local governance and keeping clinicians right at the centre.
For Australia, the timing couldn’t be better. The sector is under real pressure to do more with constrained resources, and this gives Australian health services a practical, scalable way forward that respects national values of equity, privacy and safety.
I left the conference genuinely energised and more convinced than ever that Australian health IT leaders have both the mature infrastructure and the pragmatic mindset to not just adopt this stuff, but to lead responsibly on the world stage.
If you were at HIMSS or you’re already experimenting with agents in your organisation, I’d love to hear what you’re seeing.
Observations suggest that agentic agents work safest and best when integrated with existing administrative tools. As more providers or new tools come to market, this will become a real game changer.