Scoping a medical practice in a video call with Hebora before touching any tool.
The real problem in a Brussels medical practice
A medical practice loses 20-30% of its consultation slots every year to no-shows. Multiplied by 8-12 patients per day, that's easily 40-60 lost consultations per month — the equivalent of 2-3 working days with no revenue against them.
The other problem is the phone desk. A medical assistant spends 60-90 minutes per day answering identical questions: hours, INAMI conventionnement, simple appointment booking, certificate requests. That time isn't clinical — it's pure admin and it costs money.
The real point isn't to remove humans. It's to protect the doctor's and the assistant's clinical time by filtering what doesn't deserve their immediate attention.
What can be automated in a medical practice
Online booking with consultation-specific slots (long consult, short consult, follow-up, urgent slot). Systematic D-1 SMS reminders, plus same-morning reminder for afternoon appointments. Automated patient FAQ (hours, INAMI tariffs, conventionnement, parking, accessibility, the most common certificates).
Front-line filtering of urgent vs non-urgent requests, with immediate human handover the moment a sensitive keyword appears (chest pain, bleeding, high fever in a child, etc.). Post-consultation follow-up for long treatments with milestone reminders and alerts to the doctor on missed milestones.
All patient data must be processed inside a strict GDPR framework: mandatory EU hosting, signed DPAs with all providers, retention aligned with Belgian obligations (10 years for medical records), strict access control, access logging.
Step-by-step method to automate a medical practice
<strong>Step 1.</strong> Pick a booking platform (Doctolib, Doctena in Belgium, Progenda) or scope a custom module if you want to keep your data in-house.
<strong>Step 2.</strong> Define consultation types, durations and weekly slot availability. Block urgency slots for same-day callbacks.
<strong>Step 3.</strong> Activate automatic D-1 + same-day SMS reminders. The single biggest lever to break no-shows.
<strong>Step 4.</strong> Build a structured patient FAQ and wire it to the website and/or to a front-line phone AI assistant.
<strong>Step 5.</strong> Frame escalations: any sensitive symptom routes immediately to a human, never handled by a bot.
<strong>Step 6.</strong> Measure after 8 weeks: no-shows, time saved by the assistant, patient satisfaction. Tune the configuration before scaling.
The four automations that change daily practice
<strong>1. Online booking with differentiated types and durations.</strong> Cost: Doctolib 109 €/month, Doctena 80-120 €/month, custom 2 000-5 000 € one-shot. Gain: 5-8h/week of phone-desk time saved.
<strong>2. D-1 + same-day SMS reminders.</strong> Cost: 30-60 €/month. Gain: -50% on no-shows on average (EU healthcare studies).
<strong>3. Automated email confirmations with appointment recap.</strong> Cost: bundled with the booking platform. Gain: -30% errors on date and time, professional image that reassures patients.
<strong>4. Post-consultation follow-up for chronic treatments.</strong> Cost: bundled in the medical software or 60 €/month externally. Gain: better adherence, fewer relapses, stronger patient retention.
What to avoid in healthcare
Never let a chatbot decide on a symptom or a dose. No exception. The medico-legal risk is too high, and the value of a human is exactly there.
Don't centralise patient data outside the EU. Microsoft, Google, AWS offer EU regions with healthcare-specific DPAs — use them. Consumer tools (consumer ChatGPT) are not GDPR-compliant for patient data.
Don't over-automate the relationship. A patient who only ever receives an SMS ends up changing practice. Automation handles the admin noise, the human stays on clinical contact.
Realistic costs and ROI for a medical practice
For a solo or duo practice, expect 100-250 €/month combined tool spend (online booking + SMS + FAQ), plus an initial scoping fee of 1 500-4 000 € depending on scope (single tool or full orchestration including patient records and INAMI billing).
ROI shows on three axes: no-shows halved (typical recovery of 800-2 000 €/month in regained consultations), secretarial time freed (5-10h/week, the equivalent of a half-time role), measurable patient satisfaction via Google and Doctolib reviews. Typical scoping payback in 2-4 months.
FAQ
Is medical practice automation GDPR-compliant?
Yes, under strict conditions. EU hosting, signed DPAs with all providers (booking platform, SMS, FAQ), retention aligned with Belgian obligations (10 years for medical records), access strictly limited with access logs. Hebora frames this before going live, never as a patch.
How do you cut medical no-shows?
Three cumulative levers. D-1 SMS (open rate >95%, the most effective). Same-morning nudge for afternoon appointments. One-click cancellation via the SMS. Combined, these three typically halve no-shows (from 25-30% to 10-15% in Brussels practices).
Does Doctolib replace bespoke automation?
For pure booking, yes. For full clinical workflow (exam reminders, treatment adherence, integration with practice software, INAMI billing), no. Doctolib is a booking tool, not a workflow orchestrator. For practices that want full integration, custom becomes relevant beyond 2 doctors or with specific integration needs.
What does it cost to automate a Brussels medical practice?
Indicative range: 100-250 €/month in tools depending on size (solo, duo, group). Hebora scoping: 1 500-4 000 € depending on scope. For most Brussels practices, total first-year budget runs 3 000-7 000 € with measurable return in 3-4 months.
Virtual front desk or a real receptionist?
Not in competition — they complement each other. The virtual one filters simple requests (hours, standard appointments, FAQ), the human welcomes in person, handles emergencies, manages complex cases. A practice that tries to replace 100% of the human with virtual loses patients.
Can AI help with diagnosis?
For diagnostic-support tools for doctors (Watson Health, Aidoc, etc.), yes in some areas (radiology, dermatology, ophthalmology). For talking directly to the patient: no, never. The boundary is clear: AI may help the doctor decide, it never decides on its own and never speaks clinically directly to the patient.
How do you integrate AI to the practice software (Medispring, Health One, Daktari)?
Depends on the software. Medispring and Health One have public APIs, integration is possible but requires technical scoping. Daktari is more closed — integration usually goes through Medibooking or a third-party connector. Hebora frames this before quoting to avoid surprises.
Should you tell patients you use AI in the practice?
Yes. At a minimum in the privacy policy displayed in the waiting room and on the website. If a front-line AI assistant answers the phone, it should identify itself as such and offer a switch to a human. The medical practice remains a place of trust — transparency about tools is non-negotiable.
Want to automate your business?
Hebora helps small and medium businesses in Brussels scope their automations before touching any tool. 30 free minutes to figure out what's actually worth automating in your context.
Talk to Hebora