AI Receptionist for Doctolib Medical Practices: Full Guide

Why Doctolib is not enough alone

Doctolib solved a real problem: patients can now book online 24/7 without calling. But phone volume did not disappear. It stayed roughly stable at 40-60 calls per day for a solo GP, and much higher for group practices. Why? Older patients still prefer to call. Anxious symptoms make a phone call feel safer than a form. Complex requests - reschedules with a specific practitioner, insurance questions, pre-visit anxieties - are still phone-first.

The result: Doctolib reduced booking friction for the 30-50% of patients who prefer online, but the phone lines carry the rest - and reception is still overwhelmed at 8h-10h and lunch.

How an AI receptionist plugs into Doctolib

IT Habit connects to Doctolib via Google Calendar (which the Doctolib agenda writes to and reads from) or via direct API integration on Premium and VIP plans. When a patient calls, the AI checks live availability in Doctolib, books the slot, and the booking appears in the practitioner's Doctolib agenda within seconds - identical to a booking made through the Doctolib website.

Critically, the AI respects Doctolib's slot rules: consultation duration by motive, per-practitioner availability, specialty-based routing, and blocked periods. No double-booking is possible.

Emergency triage - the non-negotiable

For a medical practice, a phone AI has to handle emergencies safely. IT Habit's medical-industry preset listens for emergency cues: chest pain, difficulty breathing, severe bleeding, suicidal ideation, unresponsive family member, and more. When any of these are detected, the agent does three things simultaneously:

1. Instructs the caller to hang up and dial 15 (SAMU) or 112 immediately.
2. SMS-alerts the practice manager with the caller number and detected cue.
3. If the practice has a nurse-triage line, transfers the call live.

The scripts are reviewed by a physician on our team - and you can customise the trigger keywords and the safety instructions per practice.

Reducing no-shows below 5%

French medical no-show rates hover around 12-15% - a huge productivity drag. IT Habit sends automatic SMS confirmation the moment the booking is made, then a 24h reminder before the appointment. Practices report no-show rates dropping to 3-5% within the first month.

The reminder can include a one-tap cancel link (which frees the slot in Doctolib) and a reschedule link. That converts the "silently absent" no-show into a rescheduled visit or a released slot.

Multilingual patient calls

Urban French practices - Marseille, Paris, Lille - regularly receive calls in Arabic, English, Portuguese, Mandarin, Turkish. A monolingual receptionist can only stall these calls. IT Habit detects the caller's language from the first sentence and handles the booking natively. Confirmation SMS goes out in the caller's language. That means fewer no-shows and fewer misunderstood appointments.

GDPR and HDS: what you need to know

IT Habit is GDPR-compliant, EU-hosted (France, Germany, Ireland), and signs a DPA. No patient data is used to train AI models - this is written into the contract. For practices requiring HDS (Hébergement de Données de Santé), we route through an HDS-certified sub-processor available on the enterprise tier.

Call recordings are optional; you can disable them entirely, or set a short retention period (30 days is typical). Transcripts are available in the dashboard with role-based access control - only the practice owner and designated admins can see full transcripts.

Deployment: 48 hours for a solo GP

Solo GP or dental practice: 48 hours from the discovery call to answering patient calls. The setup includes: voice selection (calm, warm, professional medical tone), greeting and consent script, Doctolib calendar sync, emergency triage rules, no-show reminder templates, and pricing/insurance FAQ. You spend about 2h total on the discovery call, test calls, and dashboard walkthrough.

Group practices with 3-10 practitioners: 3-5 days. VIP-tier custom PMS integrations (Maiia, KelDoc): 1-2 weeks depending on scope.

Read next