• AWS-native AI integration · ships in 6–10 weeks

Dental DSO consolidates front desk overnight

  • Voice AI · Dental
  • Three-location dental group — ~28 chairs, Dentrix Ascend, ~$5.5M collections

The problem

Front-desk staff at each location were losing ~3.5 hours/day to routine booking, confirmation, and re-care recall calls. After-hours callers got generic voicemail; the DSO's marketing team estimated 35–40% of after-hours new-patient inquiries never re-called.

Our approach

What we built

  • Inbound voice agent handling new-patient intake, existing-patient booking, prescription refill requests routed to clinical, insurance verification, and emergency triage.
  • Real-time Dentrix Ascend integration for live scheduling — not message-taking.
  • Per-location voice persona so the agent says the right practice name, not a generic greeting.
  • Strict emergency triage rubric: avulsed tooth + pain + swelling + fever routes to the on-call line, no exceptions. Weekly synthetic red-team calls verify the rubric.
  • Monthly business review with each office manager: conversion, call length, top 10 transcripts to spot-check.

Caveats

HIPAA compliance is non-negotiable. Vapi charges $1,000/month for HIPAA Zero Data Retention; we bake this cost into pricing. One front-desk team was sensitive about the agent “taking their job”; we positioned it explicitly as overflow and after-hours coverage. Insurance verification accuracy peaked at ~85% — we route low-confidence verifications to a human callback rather than guessing.

Stack. Vapi in HIPAA mode · AWS Bedrock Claude · Dentrix Ascend API · per-location persona · Langfuse evals.

Outcome

After-hours new-patient booking rose from 0% to ~18%. Front-desk time saved ~2 hours per location per day, redirected to in-office patient experience and recall outreach. Zero emergency misroutes in 90 days.

18%
after-hours new-patient bookings
0
emergency misroutes in 90 days
~3 mo
payback across 3 locations