Approved Operations for medical practices, HIPAA-aware by design
PHI-minimized scheduling, post-visit follow-up, lapsed-patient win-back. Scheduling and engagement signals only. No clinical notes, labs, or imaging. Owner-approved before anything sends.
Healthcare track in private beta. Posture statements verified 2026-05-28. See full posture →
HIPAA posture: what we do, what we don’t claim
Healthcare track in private beta. BAA available. Zero-egress local-inference path via Ollama default. SOC 2 Type I in progress. Posture statements verified 2026-05-28.
-
BAA availability & signing process
Business Associate Agreement available on request. Required before any PHI flows into the platform. Process: discovery call → scope confirmation → BAA execution → restricted-tenant provisioning → onboarding. No covered-entity pilot is live yet. Healthcare track is in private beta.
[VERIFY] Confirm BAA template language with healthcare counsel before first pilot signature.
-
Zero-egress local-inference path (Ollama default)
For healthcare tenants, the default inference path is local. RelayLaunch runs a local-LLM runtime (Ollama) with GPU passthrough and routes through an OpenAI-compatible gateway (LiteLLM) with cloud providers explicitly gated off. Patient signals never leave your environment.
[VERIFY] Confirm per-tenant LiteLLM model allow-list configuration before each healthcare go-live; capture diff in customer onboarding folder.
-
PHI minimization + redaction layer
Minimum-necessary by default. RelayLaunch ingests scheduling and engagement signals only (appointment history, recall intervals, no-show flags). It does NOT ingest clinical notes, lab results, imaging, diagnoses, or treatment plans. A pattern-matching redaction layer runs at the gateway before any model sees free-text content.
[VERIFY] Add PHI regex patterns (SSN, MRN, DOB, NPI) to LiteLLM guardrails block and confirm Langfuse telemetry routing for healthcare tenants before first pilot.
-
Audit-trail retention policy
Every AI action and owner approval is logged with an immutable trail: actor, timestamp, action, target. Every AI action is logged with reasoning, confidence scoring, and priority checks, producing the evidence record. Retention defaults to 6 years to align with HIPAA documentation expectations.
[VERIFY] Implement automated 6-year retention + WORM-style backup for audit-trail tables before first healthcare pilot crosses 90 days.
Where your revenue is leaking
Patient no-shows drain your schedule
~23% no-show rate across primary care benchmarks
Each empty slot costs $150-$500 in lost revenue and wasted staff time. A 10-provider practice with a 23% no-show rate loses meaningful annual revenue. Approved Operations stages confirmation outreach and owner-approved backfill candidates.
Post-visit follow-up falls through the cracks
Keep patients on track without manual recall list overhead
Your front desk handles check-ins, insurance, and phones. Follow-up calls get deprioritized. Patients who needed a 6-month check-up or PT follow-up never schedule. Approved Operations stages those reminders in the Daily Revenue Brief.
Cancelled slots stay empty
$150-$500 lost per unfilled cancellation
A patient cancels at 4 PM for tomorrow’s 9 AM slot. Your staff scrambles through waitlists. By the time someone answers, the slot is wasted. Approved Operations stages waitlist outreach for owner approval within minutes.
How Approved Operations sits next to your EHR. Not inside it.
RelayLaunch is intentionally read-only against scheduling data and never writes back to clinical records. Integration is sequenced by adoption and BAA scope. Healthcare track in private beta. Posture statements verified 2026-05-28.
- Day 1 · Available now
Scheduling-data import (read-only)
Flat-file CSV export from your scheduler, or read-only API where the EHR supports it. Approved Operations builds the Daily Revenue Brief from appointment history, recall intervals, and engagement flags only. No clinical fields touched.
- In-flight · First healthcare pilot
Athenahealth + DrChrono (FHIR-friendly)
Direct FHIR connection for appointment + patient-engagement endpoints. Read-only. PHI redaction enforced at the gateway before any field crosses into the inference path.
- Roadmap · Post-beta
Epic + eClinicalWorks (certified-partner path)
Epic and eClinicalWorks integrations require certified-partner status for production use. Sequenced after SOC 2 Type I attestation and the first paying healthcare pilot reaches 90 days.
- Never · By design
Write-back to clinical records
Approved Operations does not write to clinical notes, diagnoses, problem lists, medication records, or imaging. Communications-only write-back (SMS/email confirmations) is the maximum scope.
Your Monday morning, PHI-minimized and owner-approved
Daily Revenue Brief arrives
Top of the queue: tomorrow’s 10am new-patient consult open + lapsed PT patient (no follow-up in 45 days). Plus a post-visit instruction reminder for today’s follow-up and a 2pm cancellation ready to backfill from the priority waitlist. Each item is pre-written.
Approve in 30 seconds
Tap approve on the 10am consult outreach, the PT win-back, the post-visit instruction message, and the 2pm waitlist backfill. Adjust wording. Skip the rest. Done before the first patient checks in.
Focus on patient care, not admin
Approved actions execute automatically. Reminders land at optimal times. Waitlist patients fill the 2pm gap. Lapsed-patient sequences run. Your front desk focuses on the patients in front of them.
See what came back
End-of-day summary: 10am consult confirmed ($425), PT eval rebooked ($280), 2pm cancellation backfilled from waitlist ($310), 2 post-visit follow-ups returned for next-step appointments. Projected weekly recovery: $5,800.
The math
Proof period starts at $0. Starter is flat $149/mo if kept, so the math stays predictable after one $150-$500 slot is saved.
*Projections based on average medical no-show rates and lapsed-patient win-back benchmarks. Individual results vary by specialty and payer mix. Healthcare track in private beta. Posture statements verified 2026-05-28.
Is AI recommending your practice?
ChatGPT local search grew from 6% to 45% in one year. Find out if AI assistants can find your medical practice, and what to fix if they can't.
Take the Free AI Readiness Score →Start with a proof period. Keep it for a flat $149/mo.
Start with a $0 proof period while RelayLaunch proves the recovery workflow. Starter is a predictable flat $149/mo if kept. BAA reviewed before any PHI flows. Healthcare track in private beta. Posture statements verified 2026-05-28.
Slot Rescue math
What are no-shows actually costing you?
Service businesses commonly cite no-show rates around ~5-25% of bookings (industry-reported range; varies by vertical). Move the sliders to see the revenue leaving your calendar every week.
$1,800
Weekly leak
$93,600
Annual leak
$90,012
Net after Pro ($299/mo)
$81,612
Net after Team ($999/mo)
Assumes every no-show is one lost ticket. Slot Rescue targets a ~50% projected recovery (a modeled target); individual results vary by vertical and offer quality. Net figures show annual leak minus the listed annual subscription.