All case studies
Voice AI/Healthcare

Reducing missed calls across 28 clinics with a voice agent

A Japanese-language phone agent that books, reschedules and cancels appointments directly in the clinic reservation system, answers common questions and passes clinical matters to staff.

of calls handled without staff
52%
of calls handled without staff
missed-call rate
20% → 3%
missed-call rate
after-hours bookings per month
+600
after-hours bookings per month

The challenge

Most patients still book by phone, and reception teams were overwhelmed at opening time and over lunch. One in five calls went unanswered, and each missed call was often a lost booking.

The agent had to speak natural, polite Japanese, understand names and dates said in many different ways, work with the reservation system in real time, and recognise when a caller needed a person, especially for anything clinical.

Our approach

  1. 01

    Latency is the product

    On a phone call, pauses feel broken. We engineered the pipeline for fast turn-taking and let callers interrupt naturally.

  2. 02

    Clear safety boundaries

    The agent handles scheduling and general questions. Anything clinical, urgent or sensitive is transferred to staff with a summary.

What we built

Real-time voice pipeline

Streaming Japanese speech recognition, reasoning and speech synthesis with barge-in support over standard telephony.

Scheduling integration

Live availability, booking, rescheduling and cancellation in the clinic reservation system.

Escalation & transfer

Warm transfer to clinic staff with a caller summary, plus urgent-case detection.

Call analytics

Transcripts, outcomes and a dashboard per clinic for reception managers.

How it flows

  1. Phone call
  2. Speech to text
  3. Agent reasoning
  4. Reservation system
  5. Speech
  6. Book or transfer

How we measured it

Over 400 scripted test calls, including regional accents, elderly callers, noisy lines and edge cases, run before every release. Human reviewers grade a daily sample of real calls.

Results

  • 52% of calls are fully handled by the agent.
  • The missed-call rate fell from 20% to 3%.
  • About 600 additional bookings a month now come in outside opening hours.
  • Reception staff spend more time with patients in the clinic.

Timeline

  1. Weeks 1–2

    Discovery

    Call recording analysis, safety boundaries, reservation system access

  2. Weeks 3–4

    Prototype

    Booking flow on a test line, latency tuning

  3. Weeks 5–8

    Build

    Full call flows, escalation, analytics, telephony

  4. Weeks 9–10

    Rollout

    3 pilot clinics, then all 28

Contact

Have an AI product in mind?

Share what you would like to build, and we will help turn the idea into a production-ready product.

No technical specification required.

What happens next

  1. 1

    Share your idea

    Describe the problem, the users and the outcome you need.

  2. 2

    Discovery call

    We review your requirements together and clarify scope and constraints.

  3. 3

    Proposal

    You receive a clear scope, technical approach and delivery plan.