Via Médica serves patients 24 hours a day. The front desk doesn't.
Via Médica — Booking, rescheduling, exam quotes and prep instructions — handled by an AI agent on official WhatsApp, with the team stepping in only for cases that need a person.
+50,000
conversations handled by the AI since launch
85%
resolved with no human intervention
~1 min
average AI first response time
+1,700/mo
confirmed bookings
A front desk doesn't scale with demand
Via Médica receives thousands of contacts a month via WhatsApp: patients wanting to book, reschedule, ask for an exam quote or clear up a prep question. Every message depended on the front-desk queue — which also handles the counter and the phone.
Outside business hours, patients simply got no answer. And poorly communicated prep meant rescheduled exams: a loss for the clinic, frustration for the patient.
- Patients with no answer outside business hours
- A message queue competing with the counter and the phone
- Poorly communicated exam prep driving reschedules
- Slots lost for lack of active confirmation
An AI agent up front, engineering behind it
We designed an AI agent that handles the conversation end to end on the clinic's official WhatsApp: it understands the request, checks the calendar, books, reschedules, sends quotes and prep instructions — and hands off to the human team when the case calls for it.
Behind it, an orchestration layer built in n8n connects the agent to the clinic's systems, records every decision and feeds the metrics we review every month with the operation.
- 24/7 service — 100% of conversations started by the AI
- Autonomous booking, rescheduling and exam quotes
- Prep instructions at booking and the day before
- Active schedule confirmation and follow-up on abandoned conversations
- Handoff to humans with full context
From patient to metric, no black box.
Every conversation travels this path — and each step records data that feeds the continuous evolution of the operation.
Our own infrastructure · Docker · Redis for real time
Stable scale, month after month.
Around nine thousand conversations a month run by the AI — no queue, no overtime, no dependence on the size of the front desk.
What changed in practice.
Service coverage
Before
Business hours
After
24 hours, 7 days
First response
Before
Front-desk queue
After
~1 minute, by the AI
Schedule confirmation
Before
Manual call, when possible
After
Active and automatic, the day before
Abandoned conversation
Before
Patient lost
After
Automatic follow-up at the right moment
The whole operation becomes numbers.
~9,000
Conversations per month
stable at scale
+120k
Messages exchanged per month
835
Reminder confirmations (Jun)
41
Patients recovered via follow-up
in the first 2 weeks of the engine
What this operation taught us.
- 01
The biggest gain didn't come from responding fast — it came from never failing to respond. Coverage matters more than speed.
- 02
Prep instructions delivered at the right time reduce exam reschedules. The timing of information is engineering, not a detail.
- 03
Active confirmation turns the calendar into predictability: whoever doesn't confirm frees the slot in time to be rebooked.
- 04
The AI needs to know when to step aside: handing off complex cases with full context is what sustains the team's trust.
Your operation could be next.
Tell us how your operation works today. We'll show you where engineering delivers results first.
Book a diagnostic30 minutes to understand your operation, your bottlenecks and where there's a real opportunity for gains. No generic sales pitch.