What Healthcare Teaches Us About Leading AI Projects That Succeed

Healthcare is one of the most complex, regulated, and high-stakes industries in the world, which makes it a powerful testing ground for AI implementation. In a three-part AI Today × Healthcare Information and Management Systems Society (HIMSS) podcast series, senior leaders and practicing clinicians share what it takes to make AI work in this demanding environment — from easing physician burnout to reinventing documentation, diagnostics, and governance.

The lessons go far beyond healthcare. For project managers in any sector, these conversations offer a view into how to deliver AI projects that succeed under intense scrutiny, high expectations, and real-world constraints. For a proven framework to manage those kinds of projects, see PMI’s Cognitive Project Management in Artificial Intelligence (CPMAI)^™ certification.

1. Augment the Human Role

In healthcare, freeing clinicians from low-value tasks is one of the clearest opportunities for AI. Robert Havasy, Senior Director of Informatics Strategy at HIMSS, says the goal is to return time and attention to “communicating with patients, delivering empathy, sharing knowledge, and helping patients make decisions” — the work only humans can do.

As host Kathleen Walch frames it, this is “augmented intelligence”: technology that enhances human performance.

At University of Toledo Health, System CMIO Dr. R. Ryan Sadeghian applies this idea through custom GPT-based tools that generate patient-specific notes and flag important details for clinicians. Tailored to each specialty, they have cut documentation time by about 30% while improving quality. “We needed a solution that was intelligent, fast, and context-aware… one that could live inside or alongside our clinical system without requiring clinicians to become tech experts,” he says.

Project Manager Insight: In any industry, judge AI by how much it frees experts to focus on high-value work. If it simply moves the burden elsewhere, it’s missing the point.

2. Build Governance and Trust from the Start

Lasting AI adoption depends on trust and structure. At University of Toledo Health, System Dr. Sadeghian embeds AI within the clinical informatics department, under a layered governance model with an oversight committee and “trusted advisors” from every division. This approach keeps oversight in place and ensures each department has a voice. “AI suggestions are never final,” he explains. “They’re reviewable, editable… physicians maintain authorship and accountability.”

Gastroenterologist and healthtech advisor Dr. Lukasz Kowalczyk builds that trust well before launch. He meets one-on-one with anesthesiologists, nurses, schedulers, and other stakeholders to walk them through a simple mockup. Those early conversations reveal requirements that might otherwise be overlooked and build momentum for adoption. “I learned so much by going down and understanding each person’s job and their needs,” Dr. Kowalczyk says, “And that’s the way you really gain trust.”

Project Manager Insight: Governance safeguards against risk, and early engagement uncovers adoption barriers while there’s still time to address them. Make both central to your project plan.

3. Start with Data Readiness

Data maturity determines whether AI succeeds. Dr. Kowalczyk advises every organization to answer three questions before building: “How easily can you access data? How exclusive is it? How expensive is it to get?” These answers often define the project’s potential.

He also points to HIMSS research showing that 80% of U.S. hospital systems have “zero data maturity.” Dr. Kowalczyk stresses that effective AI strategies should start with the reality of current capabilities, not an idealized vision.

Havasy underscores another cost of low data maturity: much of the information hospitals collect is stored but never used. An average hospital generates about 50 petabytes a year, and 95% of it is looked at once, if at all. Storing unused data is not free — it requires infrastructure and management — which makes the inability to turn it into value a significant problem. “Most of that data is worthless until the moment some of it is priceless,” he says. For instance, “Sepsis detection…the early signs are extremely difficult to find. But if you miss them early, it’s nearly impossible to correct later on.”

Project Manager Insight: Competitive advantage comes from shortening the time between “worthless” and “priceless.” Whether you’re monitoring patient health, customer behavior, or supply chain performance, you need to have the tools and processes to surface the right data the moment it matters.

4. Lead Through Rapid Change

AI evolves faster than most traditional planning approaches can keep up with, which puts pressure on leaders to guide teams through uncertainty. Rapid change can unsettle even experienced professionals, making adaptability as important as technical know-how.

Havasy recommends three qualities for leading in this environment: flexibility, a willingness to avoid perfectionism, and principled leadership that respects those affected. “We’re linear thinkers in an exponential world,” he observes. “You need to respect the rapid pace of change and your ability to control it.”

Dr. Sadeghian adds that “the biggest challenge isn’t the technology, it’s the governance, the training, and change management.” Looking ahead, he predicts that “people who understand how to work with AI will replace those who do not… the future will belong to those who blend domain expertise with digital fluency.”

Project Manager Insight: In a fast-changing environment, guide your team with clear priorities, adaptable plans, and ongoing AI education. Equip them to respond with confidence as technology evolves — and as the demands on their roles change with it.

Putting the lessons to work

From hospital floors to corporate offices, the fundamentals of AI project success stay consistent: define a clear human benefit, establish governance and trust, prepare your data, and lead with adaptability. Whether your “patients” are customers, stakeholders, or end-users, these principles keep AI initiatives aligned with real-world needs and your teams focused on what matters most.

You can hear these lessons — and the full stories behind them — in our three-part AI Today × HIMSS podcast series:

• Robert Havasy on AI Healthcare Transformation
• Dr. R. Ryan Sadeghian: The Future of AI in Healthcare
• Dr. Lukasz Kowalczyk on Practical AI Adoption in Healthcare

If you want to turn these principles into a repeatable process, our CPMAI certification offers a globally recognized, AI-specific project management framework. Built on decades of leadership in professional standards, CPMAI combines proven project management practices with the specialized approaches AI initiatives demand — giving you the tools to bridge technical and business goals, reduce risk, and deliver measurable results.

Tags: Artificial Intelligence | Healthcare | Complexity | Data | Technology

Related Insights