OnceHub is Dead.
ReleasedPeregrine OS Scheduling is live across 11 partners, 38 providers, 12 Peregrine staff, and 13 clinic schedulers. In the last 30 days, 7,500 visits have been scheduled on a platform we built.
Emma is live at FCHC and ARcare for referrals. Our voice agent is handling referral intake at two partner sites today. Goshen is next — once they’re live, Emma will be in production across our entire managed services footprint.
Calendar sync is observable. Provider calendars connect through Google and Outlook, and until now, when a connection broke, we found out because a patient couldn’t book. The system now logs every connection failure, generates an alert when one happens, and surfaces an error log admins can view directly — so a broken sync becomes a ticket before it becomes a support call.
Missing visit recovery is in testing. When partner reports come in late or get updated after the fact, we need to cross-reference against what’s in Peregrine OS and reconcile the gaps. The recovery process is built and being tested against real partner data now — rolling out to end users next week.
Under the Hood
We shipped Dagster in production, the EMR visit pipeline, per-customer reconciliation infrastructure, referral standardization across partners, and a retry-storm fix on auth that was one CORS error away from ruining someone’s Tuesday at scale.
AI Clinical Notes
PilotAI clinical notes are piloting with providers. We’ve started a pilot of AI-assisted clinical note generation. This is early — the feedback from the first cohort will shape what comes next. Training guide is live in Flight Deck →
30 Days In
SurveyMarket Ops schedules every visit on this platform. Nine of them told us how it’s going after 30 days.
What users are asking for:
Recurring appointments for patients. Click a date and have it come up first in the scheduler. Next-day / previous-day arrows on the provider schedule view. Better guardrails against double-booking and MRN entry errors.
“Overall, it’s been great and so helpful! Thank you all so much! Amazing job.”
What’s Next
- Call Center v1 — The phone is still the front door to care for most of our partners’ patients — and right now, that front door runs through disconnected systems. Call Center v1 pulls it into Peregrine OS: inbound routing, caller identification, live booking from the call, and call intelligence that connects what was said to what got scheduled. This is the piece that turns Emma from a referral tool into the full voice layer of Peregrine OS.
- Data Foundation — As we pull data from more sources — partners, EMRs, pipelines, and Emma — we need infrastructure that can scale with it. Data Foundation is the work of making Peregrine OS maintainable at scale: a unified data model that doesn’t break as the footprint grows.
- Goshen → Emma live — Once Goshen is onboarded, Emma will be live across our entire managed services footprint. Every referral call answered, logged, and routed by the same system.
- From Pulse feedback — Recurring appointments and provider-schedule navigation (previous/next day arrows, click-to-open-date) are already scoped.
Reading List
And finally — 9 million views and a reminder for all of us building voice AI:
@huskistaken I swear I was faster #running
Husk asks ChatGPT voice mode to time his mile run. He stops it five seconds later. It confidently reports 10 minutes, 12 seconds, and when he pushes back, it doubles down: “I promise I didn’t sneak any extra seconds in there.” Then, when challenged: “I definitely have a time capability.”
