Patterns, anti‑patterns, and post‑mortems from agentic AI in production healthcare environments. Written by the people doing the work.
Most healthcare organizations think their next challenge is figuring out what to do with artificial intelligence. In reality, the first challenge is often understanding the technology environment they already have. Over the…
Read article →For the last thirty years, healthcare organizations have been building technology the same way many old cities built roads.One street was added because it solved a problem. Then another. Then a bridge.…
Read article →There was a time when people built entire homes around the limitations of the outhouse. Entire daily routines were shaped by it. Infrastructure was designed around it. Nobody questioned it because it…
Read article →Every healthcare conference feels the same right now. Someone walks onto a stage beneath glowing blue graphics and announces that artificial intelligence is going to transform medicine. Slides flash across the screen…
Read article →The reasons production deployment kills AI in healthcare are rarely about the model — they're about workflow gravity, override fatigue, and the feedback loop you build before week one.
Read article →Clinicians ignore your alert because it's wrong 80% of the time, not because they don't care. Three patterns we keep seeing in production CDSS deployments.
Read article →A specific cadence — discovery, deployment, measurement, iteration — that's gotten us through three EHR integrations without a slipped go-live.
Read article →On a real engagement, the EHR vendor refused to surface the API we needed. Here's the workaround that shipped, what it cost, and what we'd do differently.
Read article →Acceptance, modification, override-with-reason. Everything else is decoration. A short field guide to what to track from week one.
Read article →