Been running a small PT clinic and also writing dev tutorials on the side, so I sit in a weird middle ground where I understand the tooling but I'm also the one drowning in intake forms and scheduling conflicts at 7am.
Tried building some lightweight automations this past year. LLMs for parsing referral notes, some basic RAG stuff to pull patient history context faster. It works. Not perfectly, but well enough to matter.
What I keep running into is the gap between what AI demos promise and what actually holds up in a real workflow where you're shortstaffed and tired and just need the thing to not break.
That post a few days ago about AI vs human labor costs hits different when you're a small operation. You're not replacing anyone. You're trying to stop being the bottleneck yourself.
Curious what people here are actually deploying in small business or solo operator contexts. Not enterprise stuff. The scrappy builds. What broke, what stuck around, what you wish you'd done differently from the start.
[link] [comments]