We started Supa because behavioral health practices have been left out of the operations-software revolution. The clinical work is human and irreducibly so — but the layers around it (intake, documentation, eligibility, prior auth, claims) are still being done the way they were a decade ago. People stitching together five tools, eight tabs, and a phone call to do what should take thirty seconds.
This blog is where we'll share what we're learning as we change that.
What you can expect to read here
A few categories of writing, roughly in order of how often we'll publish them:
- Field notes from real practices. What's actually broken in behavioral health front-desk, documentation, and billing operations — pattern-matched across the practices we work with. Names redacted, lessons not.
- Product deep-dives. When we ship something on Supadesk, Supanote, or Supabill that we think you'd want to understand at a deeper level — not just what it does, but why we built it that way.
- The ambient AI thesis. Why "agentic" and "ambient" mean different things, why that distinction matters in healthcare, and where we think the technology is heading next.
- Benchmarks. Concrete numbers — claim accuracy, intake response time, documentation throughput. We'd rather over-share than over-promise.
A note on positioning
If you're new here: Supa builds ambient agents for behavioral health operations. Three of them so far:
- Supadesk answers calls, qualifies referrals, schedules appointments, and verifies benefits — across phones, web, and existing scheduling tools.
- Supanote listens, structures, and writes clinician-ready notes directly into your EHR — before the next patient walks in.
- Supabill runs the revenue cycle — claims, denials, appeals, eligibility — inside your existing PMS.
We say "ambient" instead of "AI" deliberately. The difference isn't just branding:
An AI feature is something you open. An ambient agent is something you forget is there — until you notice the work it quietly removed from your day.
For an IOP director drowning in front-desk callbacks, or a billing lead chasing denials, the distinction matters. They don't need another tool to learn. They need fewer tools to think about.
What we won't write
A few things we're not interested in publishing:
- Generic "AI is changing healthcare" pieces. There's enough of that.
- Anything we can't back with our own data or direct customer experience.
- Anonymous case studies dressed up as marketing.
If we don't have something useful and specific to say, we'll wait until we do.
Want to come along?
Two ways to follow:
- Subscribe to the RSS feed at
/blog/rss.xml— works with Feedly, Reeder, and most read-later apps. - Email Sam directly at hello@supanote.ai if there's a topic you want us to dig into.
More soon.