I just spent three days at HIMSS 2026. The conversation about AI in healthcare sounds nothing like what’s being written about it. There was little of the headline-driven speculation about AI broadly replacing healthcare workers.
The focus was a more practical question: how can AI improve capacity, efficiency, and experience across the healthcare system?
HIMSS President Hal Wolf opened the conference with a reality check. Healthcare is facing unsustainable costs, a worsening hiring shortage, and an aging workforce shrinking faster than it can be replaced.
In his keynote, CMS Administrator Dr. Mehmet Oz noted that roughly $6 billion is spent annually on provider identity verification alone.
Other speakers talked about rural hospitals and how they have it the toughest. Half are on the brink of closing. They survive on an average of 29 days of cash on hand.
This is the environment into which AI is being introduced. Not a system with surplus workers to shed, but a system critically short on capacity.
The dominant conversation at HIMSS was not about replacing clinicians.
The use cases being deployed right now are operational. Ambient listening that cuts documentation time. Phone agents are replacing broken phone trees. Multi-agent workflows are tearing through administrative backlogs. Coding automation. Report generation. Supply chain risk modeling.
One session put it plainly: look for the boring parts of the business.
That’s where AI is winning.
Clinical use cases are emerging as well. An urban hospital system demonstrated imaging AI finding things human radiologists missed — a 14% lift in detection. A rural hospital system built a multi-agent diagnostic workflow that was four times more cost-effective than its previous process. But in every case, humans remained in the loop.
The goal isn’t to automate clinical judgment. It’s to support it.
Every successful AI deployment at HIMSS shared one thing in common.
The organization understood its own processes before buying a single tool.
A regional health center spent months cleaning data and building a governance foundation before touching a single AI use case. As one rural health system leader put it, “Data is the AI strategy.”“ A university health system described a clear platform-first strategy: leverage existing capabilities first, evaluate the market second, and build only when necessary.
The pattern was consistent. Organizations that struggled started with the tool. Organizations finding success are starting with the problem. That may sound intuitive to experienced technology leaders, but many healthcare organizations are building these capabilities in real time across clinical, operational, and IT teams.
This theme up in nearly every session.
If a tool requires opening a new application or breaking an existing workflow, it won’t be adopted. The experience has to be seamless. Accuracy has to be higher than in any other industry. In healthcare, even small amounts of friction can undermine trust and adoption.
One urban hospital system measured AI adoption as the percentage of encounters where clinicians could have used AI tools and actually did. That’s a ruthless metric. It makes workflow gaps and usability issues immediately visible.
The pattern is clear: integration comes first. Bring AI into the systems clinicians already live in. Not the other way around.
Many organizations at HIMSS are still defining their AI governance model in practice.
One urban academic medical center shared its risk-based triage model for evaluating every AI tool before deployment. Key questions included; Does it touch patient data? Could it spread misinformation? Does it impact patient safety? The framework wasgrounded in NIST’s definition of trustworthy AI. But they were candid that standards for hallucination mitigation and LLM content controls are still evolving.
The industry is writing the rules in real time.
Organizations taking a deliberate approach are putting durable controls in place: governance boards, approved technology standards, internal testing sandboxes, and formal risk triage before launch. These aren’t bureaucratic hurdles. They’re the difference between AI that sticks and AI that gets pulled after an incident.
Before one health system introduced AI tools to nurses, leaders first had to listen to the concerns on the front line.
“Am I going to be replaced?”
“Are they going to make me take on more patients?”
These are not signs of resistance; they are valid questions about workforce impact, staffing pressure, and trust.
The organizations seeing the strongest adoption are not necessarily those with the most advanced technology. They’re the ones that led with trust, transparency, and education. They trained staff on how to experiment, not just how to operate tools. They used peer groups of early adopters as internal champions. They set a clear vision from leadership before launching anything.
One session leader said something that stuck: this isn’t like implementing an EHR. The implementation playbook for AI is still being written.
The workforce isn’t the obstacle to AI in healthcare. They are the variable that determines whether it works at all.
Hal Wolf called the next 12 to 18 months a critical inflection point for the industry.
But beneath the technology decisions, governance models, and ROI cases, the underlying objective is straightforward.
A clinician who spends less time on documentation spends more time with patients. A nurse who isn’t burned out delivers better care. A patient who gets to the right place faster has a better outcome.
That was the recurring theme across HIMSS: not the tools themselves, but the experience they enable for both caregivers and patients.
Organizations investing now in data, process redesign, governance, and workforce readiness are not chasing an AI trend. They’re building the foundation for a healthcare system that works better for everyone inside it.
That is the real opportunity for AI in healthcare.
Brian Dreyer is Senior Director of Product Management at Cloudforce, where he leads with a deep commitment to human-centered design and technology-driven innovation. A seasoned product leader, Brian brings a unique blend of product management and product marketing expertise, enabling him to translate complex customer challenges into both compelling products and clear, differentiated market positioning. Brian’s diverse skill set bridges product management, product marketing, and user experience. He has led go-to-market strategies, overseen major software redesigns, and worked hands-on in user research while collaborating closely with UX design teams. This multidisciplinary approach allows him to consistently deliver user-centric products that drive customer value, accelerate adoption, and fuel long-term growth.
At Cloudforce, we believe many of the best ideas for AI innovation are already inside the building, they just need the right space to come to life. In alignment with Microsoft’s push for...
At Cloudforce, giving back is one of our core values. Through Cloudforce for Good, we put that value into action by expanding access to education, supporting our communities, and empowering the...
FabCon 2026 made one thing clear: Microsoft is bringing its data stack into one platform, built for AI. This year's conference drew over 8,000 attendees to Atlanta, and for the first time,...
