The data is stark: A 2025 time-motion study from Vanderbilt University School of Medicine, published in The Journal for Healthcare Quality, found that registered nurses spent just 34% of their time on direct patient care, while 38% was spent on indirect tasks, primarily documentation.
Margins are under pressure. Clinicians are burning out. We’re asking staff to deliver excellent care while the infrastructure works against them.
AI Bolted Onto a Fragmented Environment Doesn’t Work
There’s been enormous enthusiasm for artificial intelligence in healthcare. Unfortunately, I see organizations make the same mistake repeatedly: They layer AI onto a fragmented foundation and expect it to fix the underlying problem. It doesn’t.
Large language models are powerful, but they are not a system of governance. They don’t enforce ownership. They don’t connect systems of record or complete workflows. An AI copilot on top of hundreds of disconnected applications just exposes fragmentation.
What the Right Foundation for AI Looks Like
For AI to deliver real value in healthcare operations, it needs a unified foundation: a shared operational system of record embedded directly into the clinical environment, including the electronic health records system, so care teams can raise requests as part of their normal workflow without breaking focus.
DISCOVER: Learn how healthcare organizations are unlocking AI’s potential.
When that foundation exists, AI agents can do real work: triaging requests, routing them to the right team and keeping tasks moving without manual intervention.
Consider something as simple as broken equipment on a hospital floor. Today, a nurse calls the biomed department, waits on hold, explains the problem and hopes it gets tracked. With a unified operational platform, she flags it in the system she’s already working in. It’s automatically routed with the context the support team needs. The issue is resolved with clear ownership and real-time status. Every action is captured because the platform recorded it as work happened, not because someone filed a report.
The Stakes To Make Agentic AI Work in Healthcare Are Real
Clinicians trained to heal are spending their days navigating broken systems. That’s not an AI problem; it’s a systems design problem. And when support services move at the speed clinicians need, the effects compound: fewer care disruptions, reduced burnout, and operating margins protected from the waste of manual coordination.
READ MORE: Build a resilient AI ecosystem in healthcare beyond compliance.
My husband didn’t go to medical school to fight with fax machines. Nobody reading this went into healthcare to watch their teams drown in process. The question is whether we keep layering tools on top of a broken foundation or finally build the right one.
The organizations getting this right aren't starting with AI. They're starting with the foundation AI needs to finish work: one operational platform, shared across departments, embedded in the clinical environment.
