Case study / Healthcare / AI Telephone Answer Service
Every call answered, 24 hours a day, across three practices.
A growing dental group was missing more than a third of inbound calls during peak hours. An AI receptionist now answers instantly, books and reschedules appointments, and hands urgent cases to the on-call team.
- Industry
- Healthcare
- Client
- Independent dental group
- Scale
- 3 practices · 14 chairs · 6 front-desk staff
- Timeline
- 5 weeks from discovery to live
- Solution
- AI Telephone Answer Service
2%
calls missed
down from 38% at peak
+41
new-patient bookings per month
22 hrs
front-desk time returned each week
27%
of new bookings made out of hours
01 / The challenge
Where the time was going.
Front-desk teams were doing two jobs at once: greeting patients in the room and answering a phone that never stopped. Between 8am and 10am and again over lunch, calls rolled to voicemail. Most of those callers were new patients who simply rang the next practice on the list.
- 0138% of calls unanswered during the busiest 4 hours of the day
- 02Voicemails returned on average 3.5 hours later, often too late to win the booking
- 03No coverage evenings or weekends, when many working patients try to book
- 04Reception staff regularly interrupted mid-check-in to take routine calls
- 05No reliable record of why people called or how many enquiries were lost
02 / The solution
How the workflow runs.
We deployed an AI receptionist on the group's existing phone numbers. It answers every call in under two rings, handles the routine work end to end, and steps aside for anything a human should own.
- I
Answer and understand
The assistant greets the caller by practice name, listens to the request in natural language, and identifies whether it is a booking, a change, a question, or an emergency.
- II
Check live availability
It reads real-time diary availability from the practice management system for the right clinician and appointment type, then offers two or three concrete slots.
- III
Book, confirm, and record
The appointment is written straight into the diary, the patient receives an SMS confirmation, and a call summary is logged against their record.
- IV
Triage and escalate
Pain, swelling, or trauma keywords trigger the emergency protocol: the caller is connected to the on-call clinician and the practice is notified.
- V
Hand off with context
Complex treatment or billing questions are transferred to reception with a one-line summary, so the patient never repeats themselves.
Human-led by design
Guardrails that kept people in control
- Never gives clinical advice. Symptom questions are routed to a clinician.
- Insurance and pricing answers come only from an approved knowledge base that the practice manager edits.
- Every call is transcribed and reviewable, with a daily digest of anything the assistant could not resolve.
- Patients can say “speak to a person” at any point and are transferred immediately.
Stack
- Twilio Voice
- OpenAI Realtime
- Practice management API
- SMS confirmations
- Slack alerts
Timeline
- Week 1
Discovery
Listened to 200 recorded calls, mapped the top 12 call reasons, and agreed escalation rules with clinical leads.
- Weeks 2–3
Build
Integrated the diary, wrote the approved knowledge base, and rehearsed edge cases with reception staff.
- Week 4
Pilot
Went live on one practice during lunch hours only, then extended to the full day after a review.
- Week 5
Rollout
Enabled all three practices with 24/7 cover and a weekly quality review.
03 / The results
Before and after.
Measured over the 90 days after go-live against the 90 days before. Figures are rounded and client details anonymised.
| Measure | Before | After |
|---|---|---|
| Calls missed at peak | 38% | 2% |
| Average time to answer | 47 seconds | Under 2 rings |
| New-patient bookings per month | 96 | 137 |
| Voicemails requiring callback | ~310 per month | ~20 per month |
| Front-desk hours on the phone | ~34 hrs per week | ~12 hrs per week |
What the team noticed
- Reception now focuses on patients in the building, and check-in queues have visibly shortened.
- The practice manager sees a daily call log with reasons and outcomes, which feeds staffing decisions.
- Clinicians report fewer gaps in the diary because cancellations are refilled the same day.
What comes next
- Outbound recall reminders for patients overdue a hygiene visit
- WhatsApp booking using the same availability logic
- Multilingual answering for the practice serving a largely Polish-speaking community
Want this for your business?
We start with one workflow, measure it properly, and expand only when the numbers say so.