What small hospitals teach big systems about AI adoption
Running micro hospitals taught me that small teams cannot hide a bad workflow. The same is true of AI, and it is why so many pilots stall.


I helped run two small hospitals in North Texas that became micro hospitals. A small hospital teaches you something big systems can avoid for years: a bad workflow has nowhere to hide.
When you have one nurse at the front and one physician on the floor, every extra click lands on a real person, at a real hour. If a process does not work at 3 a.m. with a short team, it does not work.
The same test applies to AI
Most hospital AI pilots I see look great in the demo and quietly die in the workflow. Small hospitals would have caught that on day one. Three questions from that world apply to any AI deployment:
- Who does the extra step? If the answer is "the nurse," you have not removed work. You have moved it.
- What happens on the worst night? Test the tool when the team is short and the department is full, not on a calm Tuesday.
- Who owns the outcome when it is wrong? If no one can answer that, the tool is not ready.
That is the core of my keynote, AI That Survives the Workflow. You can read the full operator lessons on Micro Hospital Guide, or bring the talk to your team.
