Artificial intelligence has become the defining conversation in healthcare. Every conference agenda, executive strategy session, and boardroom discussion includes AI as a top priority. Yet despite the excitement, most health systems are still asking the same question: How do we move from promising pilots to meaningful transformation?
In a recent episode of The Big Unlock podcast, Dr. Anne Snowdon, Scientific Director and CEO of SCAN Health and Chief Scientific Research Officer at HIMSS, offered a refreshingly pragmatic perspective on this challenge. Rather than focusing on the latest AI models or breakthrough algorithms, she shifted the conversation to something far more fundamental: the systems, evidence, and infrastructure required for AI to create lasting value in healthcare.
Drawing from decades of experience spanning nursing, digital health research, healthcare supply chains, and strategy, Dr. Snowdon argues that healthcare’s future won’t be determined by who adopts AI first. It will be determined by who builds connected, evidence-based ecosystems capable of supporting AI safely, responsibly, and at scale.
She identified three key barriers preventing health systems from scaling AI transformation:
- AI Continually Learns and Adapts, Making It a Moving Target
- Limited AI Education and Literacy Across the Healthcare Workforce
- Healthcare Is Still in the Early Stages of Building Evidence for AI’s ROI
Her perspective is particularly valuable because it moves beyond technology hype. Instead, she challenges healthcare leaders to think differently about evidence, workforce readiness, patient empowerment, and the often-overlooked infrastructure that makes modern healthcare possible.
Listen to the full conversation
Healthcare is Still in the AI Pilot Phase
One of Dr. Snowdon’s most important observations is that healthcare is still remarkably early in its AI journey. While organizations across North America, Europe, and Asia-Pacific are actively experimenting with AI, most deployments remain limited to pilots and proof-of-concept initiatives. Ambient documentation, clinical note generation, and workflow automation have demonstrated encouraging results, but enterprise-wide transformation remains elusive.
This assessment mirrors what many healthcare executives have shared on The Big Unlock. Organizations recognize AI’s enormous potential, but few have successfully scaled these innovations across entire health systems. Unlike many industries, healthcare cannot afford to deploy technology based solely on enthusiasm.
Patient safety demands something stronger. It demands evidence. Dr. Snowdon’s background as a nurse reinforces this point. Before introducing any technology into patient care, clinicians need confidence that it improves outcomes and does not introduce unintended risks. That evidence-first mindset has long guided medicine, and AI should be no exception.
AI Requires a Completely New Model of Evaluation
Traditional healthcare technologies are relatively static. A medical device or pharmaceutical treatment is tested, approved, and then deployed with relatively predictable behavior. AI is fundamentally different.
Machine learning models continuously evolve. Their outputs may change as data changes. Performance can improve or degrade over time. According to Dr. Snowdon, healthcare currently lacks mature methodologies for evaluating technologies that continuously learn. Rather than treating AI implementation as a one-time procurement decision, organizations must adopt a lifecycle approach that continuously measures performance, safety, bias, and clinical value after deployment. This represents one of the biggest conceptual shifts facing healthcare leaders.
Success isn’t simply about selecting the right AI solution. It’s about building governance processes that ensure AI continues delivering value months and years after implementation. Organizations that invest only in AI technology but not in continuous evaluation will struggle to realize sustainable returns.
Healthcare Must Learn to Think in Probabilities, Not Certainties
Perhaps the most thought-provoking insight from Dr. Snowdon centers on how AI changes clinical decision-making itself. Healthcare has traditionally been built around deterministic thinking:
- Assess the patient.
- Establish a diagnosis.
- Follow an evidence-based treatment pathway.
AI introduces something different. Instead of offering certainty, AI often generates probabilities. It:
- predicts which patients may deteriorate
- estimates infection risk
- identifies individuals most likely to benefit from early intervention
That may sound like a subtle distinction, but it fundamentally changes how clinicians interact with technology. Healthcare professionals have been trained to seek definitive diagnoses. AI asks them to incorporate predictive insights into clinical judgment, often before a condition fully develops. Moving from reactive medicine toward predictive care requires not only new tools, but new ways of thinking. As Dr. Snowdon explains, this cognitive shift may prove just as significant as the technological one.
AI Literacy May Become Healthcare’s Biggest Competitive Advantage
Technology alone will never transform healthcare. People will. One of Dr. Snowdon’s strongest messages is that today’s healthcare workforce has not yet received sufficient education about what AI can and cannot do. Many clinicians still view AI as a “black box.” That uncertainty naturally creates hesitation. Without adequate education, even highly capable AI solutions risk low adoption because clinicians lack confidence in interpreting AI-generated recommendations. Future AI strategies therefore cannot focus solely on software procurement. They must also include:
- Executive education
- Physician engagement
- Nursing education
- AI governance training
- Change management
- Continuous learning programs
Healthcare organizations that invest in workforce capability alongside technology will likely achieve much higher adoption rates than those focused exclusively on implementation. In other words, AI literacy may become as important as digital maturity.
Patients Must Become the Center of the Digital Health Ecosystem
Much of today’s digital transformation still revolves around healthcare organizations: electronic health records, hospital workflows, provider productivity, and operational efficiency.
Dr. Snowdon envisions something different. She believes AI should increasingly empower patients themselves. Rather than existing as disconnected consumers of healthcare services, patients should become active participants in digitally connected ecosystems where AI helps them better understand, manage, and navigate their health while remaining seamlessly connected to trusted clinicians. This is an important distinction.
The future isn’t simply about hospitals becoming more intelligent. It’s about people becoming more connected to their own health. As wearable devices, remote monitoring, patient-facing AI assistants, and interoperable health platforms continue to mature, healthcare can shift from episodic treatment toward continuous engagement. That evolution has the potential to improve outcomes while strengthening relationships between patients and care teams.
Supply Chains May Be Healthcare’s Most Underrated Digital Asset
One of the most distinctive aspects of Dr. Snowdon’s perspective comes from her decades of research into healthcare supply chains. Supply chains rarely receive the same attention as AI, clinical decision support, or digital therapeutics. Yet she argues they represent foundational infrastructure for modern healthcare.
The right product, available at the right time, delivered to the right patient, supported by accurate data – these seemingly operational functions directly influence quality, safety, and patient outcomes. When supply chains become digitally connected, they enable greater visibility, better resource allocation, and stronger integration across clinical and operational workflows.
In many ways, AI becomes far more powerful when built on top of these connected systems rather than isolated datasets. This systems-level perspective distinguishes Dr. Snowdon’s thinking from many AI discussions that focus narrowly on algorithms instead of the infrastructure supporting them.
Why Connected Systems Matter More Than Individual AI Applications
Throughout the conversation, a consistent theme emerges. Healthcare transformation isn’t about deploying hundreds of AI tools. It’s about connecting data, workflows, clinicians, patients, and infrastructure into a cohesive ecosystem. Disconnected technologies create fragmented experiences. Connected systems create intelligent healthcare.
This is where interoperability, governance, evidence generation, workforce readiness, and digital infrastructure intersect. AI can certainly accelerate clinical documentation. It can improve scheduling and support diagnosis. But its greatest long-term impact may come from connecting previously isolated parts of healthcare into coordinated learning systems that continuously improve over time.
Final Thoughts
The healthcare industry is understandably excited about AI. But as Dr. Anne Snowdon reminds us, excitement alone won’t transform care. Real transformation requires evidence. It requires education, governance, connected infrastructure, and above all, it requires keeping patients, and not technology, at the center of every innovation.
Perhaps her most powerful contribution is reframing the conversation itself. Instead of asking, “How quickly can we deploy AI?”, healthcare leaders should be asking, “How do we build connected, evidence-based systems that allow AI to improve safely over time?” That shift in thinking may ultimately determine which organizations move beyond experimentation and create lasting value.
As health systems continue navigating the next chapter of AI adoption, Dr. Snowdon’s message is both timely and enduring: the future of healthcare will not be built by AI alone; it will be built by connected, evidence-based systems that combine technology, trusted data, empowered clinicians, and engaged patients into a single learning ecosystem.