All cases
Voice AI
MediCare Group

AI reminder calls cut no-shows by 41%

-41%
No-shows
92%
Reminders answered
11h
Staff time saved/wk

Client profile

MediCare Group runs several outpatient clinics covering primary care, diagnostics and specialist consultations, with every physician's day cut into fixed slots from early morning to evening. Scheduling still runs on the phone: patients call the front desk, and the front desk is supposed to call them back. Each site has a small admin team handling check-in, payments, paperwork and the phone at the same time, so the day before an appointment is the point where the schedule either holds or quietly falls apart.

The challenge

No-shows were costing the group thousands a week, and front-desk staff had no time to call every patient the day before.

Text reminders were ignored; a real voice that could reschedule on the spot was what patients actually responded to.

The damage never looked dramatic in the moment. A patient did not arrive at 11:20, the doctor waited, the next one came in on time, and the slot was simply gone. Repeated across every site and a full roster of specialists, that quiet leak reshaped the whole week: physicians sat idle while other patients were being told the nearest opening was two weeks out.

The front desk knew the cure was a call the day before, and knew they would never get through it. Between check-in, payments, insurance paperwork and the queue at the window, evening call rounds started at the top of tomorrow's list and stopped the moment the waiting room filled up. The names at the bottom were never dialed at all.

Text reminders moved the work rather than solving it. A message landed while the patient was at work, got read once and forgotten, and left nowhere to say that Thursday no longer worked. The people who genuinely needed a different time had no fast route to one, so they picked the easiest option: skip it and call back someday.

Why earlier attempts failed

The solution

A MyChatBot voice agent now calls every patient on tomorrow's schedule, confirms the visit in a normal conversation, and moves the appointment to a free slot before the call ends.

  1. Pull tomorrow's list Webhooks API exports the next day's appointments from the clinic system every evening, and Task Manager queues each patient for a call.
  2. Call in a human voice MyChatBot Calls SDK places the outbound call with a natural voice in the patient's language, and the Agent On/Off schedule keeps dialing inside decent hours only.
  3. Reschedule inside the call Google Calendar and Altegio give the agent live availability per doctor, so a patient who cannot make Thursday hangs up already booked into another slot.
  4. Write every call back to CRM create_crm_lead and edit_crm_lead update the patient record with the outcome, lead_chat_link attaches the recording and transcript, and the source is tagged AI-.
  5. Escalate what needs a person Hand-off Control with a stop word and silent hand-off passes clinical questions and complaints to the front desk, while Flight Control pings the team in Telegram.
  6. Retry and backfill slots Follow-ups redial unanswered numbers at a different time of day, Lead Processing labels mark confirmed, moved, cancelled or needs a human, and freed slots go out as outbound calls to patients waiting for an earlier date.
Knowledge Base: The Knowledge Base holds the group's Business Domain docs: clinic addresses and how to find the entrance, the specialty and doctor schedule with slot lengths, preparation rules per appointment type (fasting before blood work, referral and ID to bring, arrive earlier for a first visit), the cancellation and rescheduling policy, and the hard line the agent never crosses on the phone, which is test results, diagnoses and treatment advice.

No-shows fell by 41% and the front desk got eleven hours a week back.

What was deployed

Agents used

Voice AgentLead QualifierAppointment Booker

Channels connected

VoiceWhatsAppWeb widget

Rollout timeline

Week 1
Front-desk call recordings and clinic rules turned into Business Domain docs, plus the nightly appointment feed from the clinic system.
Week 2
Configuration Wizard battle-testing on one site and one specialty, with every recording reviewed the next morning and the instruction versioned.
Weeks 3 to 4
Rollout to all sites and specialties, evening calling window, Follow-up retries and Telegram alerts for the front desk.
Month 2
Outbound waitlist calls added, so a slot freed by a reschedule is offered to another patient the same day.

Results

MetricBeforeAfter
Missed appointmentsEmpty chairs discovered only when the patient failed to walk in-41% no-shows
Reminder pick-upTexts went out and nobody knew who had read them92% of reminder calls answered
Front-desk time on call roundsDialing squeezed between check-ins and abandoned mid-list11 hours a week back
ReschedulingPatients who could not make the time simply did not comeMoved to a free slot before the call ends
Coverage of tomorrow's scheduleOnly the names the desk reached before closingEvery patient on the list called, missed numbers retried the same day

"Patients pick up because it sounds like a person who can actually help."

Iryna Sydorenko
Patient Care Manager

What comes next

Recall campaigns are next: annual check-ups, post-treatment follow-up visits and patients who have not been in for a year, all dialed by the same agent. WhatsApp is being added as a second lane for people who prefer to confirm in writing.

Ship something like this

Set up your first AI agent in 10 minutes. No credit card needed.

Start free trial
See all cases