How AI Reaches Organizations as Option-Rich Infrastructure An executive team I spoke with recently had a familiar problem: they weren’t just exploring AI—they were trying to “AI-ify” multiple functions in their workflow. But the market has exploded so quickly that there simply aren’t enough hours in the day to properly diligence every option. Every week, a new assistant, copilot, or “smart” feature promises to solve a specific problem. Most are shallow, non performant or wor

Truly Successful “AI Projects” in Non-Acute Care Result in Purchase of Systems-of-Record – NOT Chaotic Widget Collections In engineering terms, treating AI as a crosscutting “project” has been the unfortunate norm in non-acute care– necessitated by the lack of truly AI-infrastructured EMRs in the marketplace. Driven by marketing from legacy providers that have not upgraded old tech stacks, buyers have experienced AI so far as shadow layer of services, data flows and governanc

The real product is AI infrastructure. EMRs that forget that become technical debt. Operators know when a system is working for them and when they’re just paying rent on somebody else’s technical debt. In 2026, that’s the real line between AI “platforms” that will survive and those that will disappear. In a recent post, I argued that in healthcare tech, “anything is possible” often means “nothing will scale.” We’ve all seen the pattern: dazzling demos, endless promises of c