MSN, FACHE, AIGP | Harvard Senior Executive Fellow
Healthcare Operations Leader | AI Adoption Strategist | Clinical Co-Designer
Engineering the human and operational infrastructure AI needs to succeed in healthcare
Get in Touch
AMAZON
#1 BESTSELLER
Healthcare organizations are deploying AI faster than governance can keep pace. When implementations accelerate without infrastructure, the gap doesn't manifest as one organizational problem—it surfaces differently across your entire leadership team.
The OJC Operating System is how you engineer the human and operational infrastructure AI needs to succeed — one connected system that turns AI capability into realized value. The compass defines what your organization refuses to lose; the mechanisms make it real.
Any organization can copy a single exercise. The return doesn’t come from the pieces — it comes from how they connect. Each mechanism holds up a point the others depend on, so one loop compounds toward value instead of drift. Remove one, and the loop leaks.
And it starts with identity, not workflow.
When AI enters real work, the first thing that moves is whether people still feel needed. Engaged people keep exercising judgment and correcting the machine; people who feel replaceable go quiet — and the system learns from the silence. The same technology, dropped into engagement or into fear, produces opposite outcomes.
People locate where their judgment, presence, and growth stay indispensable — instead of fearing displacement.
Leaders commit, out loud, to what they’ll protect — so speaking up changes something real.
A shared language for judgment, so it stays visible instead of private.
Turns one person’s catch into a shared guardrail — so the organization learns and the AI improves.
The same system that protects judgment lets you measure it. I help you build the two layers most programs skip — Leading Indicators that surface drift before outcomes move, while it’s still cheap to fix, and Value Realization that proves the return against a baseline: harm prevented, expertise retained, your best people kept, and value you can safely scale. The method is the engagement — what you keep is proof you can put in front of your board.
Melinda deHoll, MSN, FACHE, AIGP, is a healthcare executive with deep clinical experience and more than three decades of responsibility for large-scale healthcare operations, strategy, training, and governance at a national healthcare organization.
Advanced leadership training in complex systems management and organizational transformation.
A master's-prepared nurse, Melinda has worked across clinical practice, operations, training, leadership development, and executive functions. She has overseen enterprise training portfolios spanning hundreds of facilities and held accountability for systems where clinical judgment, safety, training, workforce management, and human performance intersect at scale.
She co-designed an enterprise-scale clinical AI prototype through the Veterans Health Administration's competitive innovation pipeline, advancing from VHA/MIT Hacking Medicine to VHA Make-a-thon and ultimately to development collaboration with Microsoft through VHA Venture Studio.
As a certified AI Governance Professional (AIGP), she brings validated expertise in AI governance frameworks, risk management, and responsible AI deployment.
Her work on AI adoption and human factors in healthcare has been published by the American College of Healthcare Executives (ACHE).
Melinda understands what happens after vendors leave and pilots end—the operational reality of governing AI systems that influence clinical judgment under time pressure, at scale, across variability. Her frameworks emerge from that lived experience.
Engineering the Human Infrastructure AI Needs to Succeed
The technology works. What fails is the human infrastructure—the frameworks, governance, and cultural mechanisms that determine whether AI adoption protects patients or introduces new forms of harm.
This book provides the missing infrastructure: practical frameworks (OJC, AMOE, Trust Contract, Friday Proof) that make AI governance operational, not aspirational. Written for leaders who must build what doesn't yet exist—before speed, pressure, and drift make the choice for them.
Published by Echelon Press | February 2026
I work with leadership teams navigating AI adoption—from board presentations to implementation consulting to team learning sessions.
If this book changed how you see your system, you're already doing the hardest part.
You're not doing it alone.
The pace of change right now is unlike anything we've faced before. AI systems are learning. Healthcare systems are learning. And we—the leaders, clinicians, and operators building the infrastructure that keeps both aligned—are learning too.
This transformational technology is rapidly changing and emerging. No one has completely figured this out. There are no experts in preventing problems that didn't exist three years ago.
Instead we have each other.
What I've shared in this book came from hundreds of conversations with leaders navigating these same tensions—people willing to name what they were seeing, share what wasn't working, and build together when no playbook existed.
That's how we succeed: not by waiting until we're certain, but by learning faster than the systems we're trying to govern.
The frameworks in this book synthesize decades of safety science—High Reliability Organization principles, psychological safety research, systems learning theory, and AI governance models—applied specifically to clinical AI adoption.
The principles are proven. Every mechanism builds on concepts that have prevented harm in healthcare and other high-reliability industries for years. These frameworks synthesize proven safety principles applied to clinical AI adoption. They're informed by direct implementation work in clinical AI environments where these approaches were applied in real workflows.
Formal multi-site validation will take years—harm is occurring now. Early adopters who build this infrastructure now gain strategic advantage while others are still debating whether it's necessary.
These frameworks aren't meant to be final. If you implement elements in your system and discover what works differently in your context, I want to learn from you. If you find adaptations that serve your organization better, share them. The field will only succeed if we learn faster together than any of us could alone.
I'm always learning from leaders doing this work—and happy to help where I can.
If you want to think through something, share what you're building, or simply continue the conversation, reach out: