Is 2025 the year AI truly comes to life in healthcare? If HIMSS 2025, held last week at the Venetian Convention and Expo Center in Las Vegas, is any indication, the answer is a resounding yes. I spent the week immersed in sessions, navigating crowded exhibit halls, and engaging with industry peers. AI wasn’t just a topic—it was omnipresent. From the HIMSS bookstore stocked with more AI and machine learning titles than ever before (including mine 😉) to the 28,000+ attendees buzzing about pilots and proofs-of-concept (POCs), the conference showcased an industry on the cusp of transformation and lasting change.
Here’s my unvarnished take on health AI at HIMSS 2025—celebrating the great, unpacking the good, and calling out the bad.
1. AI Everywhere: Momentum Meets Maturity
AI was on everyone’s lips, woven into keynotes, panels, and casual hallway conversations. Sessions and exhibits highlighted tangible progress: ambient/voice AI, generative AI-powered chatbots, and document intelligence dominated the discourse. Compared to past years, I noticed a marked uptick in discussions—and vendor pitches—around AI in medical imaging and traditional machine learning (ML) use cases.
Generative AI is the rising tide lifting all ships, enabling use cases that blend cutting-edge innovation with tried-and-true predictive analytics. Ambient/document/conversational AI are on track to augmenting & amplifying clinical & administrative workflows. At Velatura, we’re seeing similar promise in our own AI First efforts, where consent, AI governance, and structured outcomes make value creation measurable and repeatable.
My take? If this trend holds, 2025 could mark a tipping point where AI moves from experimentation to enterprise-wide adoption, delivering real ROI for providers and patients alike.
2. AI Washing: Hype vs. Reality
Not all that glitters is gold, and HIMSS 2025 had its share of “AI washing”—when vendors slap an “AI-powered” label on products that are little more than basic algorithms or API integrations with public foundation models like ChatGPT. I lost count of the breathless pitches that, upon closer scrutiny, revealed more sizzle than substance. One vendor boasted an “AI chatbot” that turned out to be a glorified rules engine; another’s “AI platform” was just a shiny wrapper around a third-party API with no custom training or healthcare-specific tuning or clinician/patient feedback loops.
This isn’t just misleading—it’s damaging. AI washing erodes trust, overpromises capabilities, and leaves buyers with tools that can’t answer critical questions about privacy, transparency, or clinical relevance. For an industry where every decision impacts lives, this is a red flag. My take? Tech buyers and healthcare leaders: caveat emptor. Dig into the specs, ask hard questions about model training and data provenance, and don’t fall for upsells that lack substance.
3. AI Governance: The Bedrock of Trust
AI governance—spanning transparency, interpretability, explainability, privacy, and trusted data use—was a recurring theme. The expo floor reflected this priority, with the big three cloud providers (AWS, Microsoft, Google), John Snow Labs, and consultancies touting ISO 42001/NIST-compliant solutions and governance services. Yet, maturity levels varied wildly. Few vendors or buyers showcased comprehensive programs addressing AI’s diverse applications—clinical diagnostics, administrative automation, and beyond.This gap is a challenge and an opportunity. At Velatura, we’re starting small but smart, focusing governance efforts on high-impact workflows and decisions at the point of use/impact.
My take? governance isn’t optional—it’s foundational. We must shift from obsessing over data quality and model creation to ensuring data fidelity and responsible use. Trust hinges on disclosure—patients and clinicians deserve to know when, where, and how AI shapes decisions. HIMSS 2025 reinforced that without governance, even the most brilliant AI risks becoming a liability.
In Summary: From Pilots to Production
HIMSS 2025 felt like a turning point—healthcare enterprises are moving beyond AI pilots to production-ready capabilities. The appetite is undeniable for ambient AI, conversational bots, contextually intelligent agents, and document intelligence. ML-based predictive analytics and custom foundation models are poised for significant adoption. These advancements promise to enhance patient experiences, boost operational efficiency, and unlock new frontiers in research, care delivery, and health economics.
But challenges loom. Established processes resist disruption, workforce training lags behind AI’s rapid evolution, and governance frameworks are playing catch-up.
Looking Ahead: Opportunities and Imperatives
What opportunities and challenges does this AI surge create? How do we harness creativity to serve our teams and patients?
We see AI as a catalyst—improving patient experiences through personalized care, enhancing health economics with predictive insights, and enabling prosperity by empowering patients, clinicians, and their families/communities.
The imperative is clear: we must blend innovation with accountability, training our workforce to wield AI effectively while embedding governance into every layer of our strategy.
HIMSS 2025 wasn’t just a conference—it was a call to action. As I debrief with our team, we’re doubling down on AI that’s practical, trustworthy, and human-centric. The road ahead is complex, but the potential to transform healthcare has never been greater. See you at HIMSS 2026—hopefully with even more stories of impact to share.