Job · Appointment booking
A patient asks for a slot and leaves with one booked.
Scout reads real availability through an endpoint your clinic exposes, offers the times that are actually free, and confirms the booking in the same thread. The moment the patient asks something clinical, it stops and hands the thread to a person.
عرض توضيحي ببيانات وهمية.
What it looks like
Hana Mostafa asks for a slot. Scout books one, then refuses a question.
One thread, start to finish: a request, a real availability lookup, three offered times, a confirmed booking — and a guardrail where the conversation turns clinical.
The patient asks
“Do you have a physio slot this week?” arrives on WhatsApp. The thread already carries the patient's contact record, so there is no round trip asking who is writing.
Scout reads availability
A booking-availability tool the clinic configured runs against its own diary. The call and what it returned appear in the trace; nothing about free time is guessed.
Three real times are offered
Wednesday 11:00, Thursday 18:00 or Saturday 09:30 — the slots the tool returned, offered in the clinic's own words.
The patient picks, and it is booked
“Thursday 18:00 please.” The booking is written back through the same tool, confirmed in the thread, and a reminder is scheduled for 24 hours before.
Then the question turns clinical
“Is it safe after knee surgery?” The guardrail fires before a reply is drafted. Scout says a clinician has to answer that, and does not attempt it.
Handover with the whole thread
The conversation goes to Dr Yara Selim's assistant with the booking, the tool result and every message attached. The patient keeps the appointment they just made.
The problem
The phone is busiest at exactly the hours patients are free.
Booking is not a hard conversation. It is a conversation that arrives in volume, at the wrong time, on four channels at once.
- Engaged tone at 08:30 and 19:00
- Patients call before work and after it, which is when the lines are already full. A missed call is a patient who tries the clinic down the road.
- Requests in four places, recorded in none
- WhatsApp, Instagram, the website form, a voicemail. There is no shared record, so two people can be offered the same Thursday slot.
- The same three questions, retyped
- Which clinician, which day, new patient or returning. The front desk types them out again for every request and then types the answer into the diary.
- The real fear: a bot that answers a medical question
- A clinic will accept a slow reply. It will not accept software interpreting a symptom. That is the risk that stops most clinics automating anything.
How it works
The mechanism, step by step.
Booking is a tool call wrapped in a scenario. The diary is authoritative; the model only writes the sentence around what it returned.
- The booking scenario
- A scenario you can read and edit: which clinicians can be offered, how many slots to put in front of a patient, what to ask a new patient, when to stop and fetch a person.
- Availability is your endpoint
- There is no built-in EMR or practice-management integration. The clinic exposes an HTTPS tool that returns free slots and accepts a booking; Scout calls it and quotes exactly what came back.
- No invented slots
- If the tool returns nothing, the patient is told the diary is full and the request goes to the front desk. An empty response is never filled in with a plausible time.
- Guardrails on clinical content
- Symptoms, medication, suitability, anything diagnostic. The guardrail is evaluated before the reply is sent, so the refusal is structural rather than a line of prompt.
- Handover carries context
- The assigned person opens the thread with the full history, the tool result and the booking already visible. Meta's policy requires that escalation path to exist.
- The reminder follows automatically
- A confirmed booking schedules a utility template 24 hours ahead, which the patient can reply to. That is the next job: appointment reminders.
What it costs to send
The booking is free-form. The confirmation you send later is a template.
The patient started the conversation, so the whole booking exchange sits inside the 24-hour service window. Anything you send after the window closes needs an approved template, and its category decides the price.
- The booking conversation itself
- Service message
- A message from the customer opens a 24-hour window. Meta began charging for service messages on 1 October 2026, with the first 1,000 per business phone number per month free.
- Confirmation or reminder sent later (Egypt)
- $0.0036
- Utility template — a transaction the patient agreed to. Sent when you reopen the conversation, not when the patient writes.
- Re-engaging a lapsed patient (Egypt)
- $0.0644
- “It has been a year since your last visit” is marketing, whatever it is about. Roughly eighteen times the utility rate.
Source: Meta WhatsApp per-message rate card, Egypt, retrieved 2 November 2026; category, window and service-message rules from Meta developer documentation.
Meta bills these fees, not us, and we never mark them up. BlendDesk charges for seats and AI resolutions.
Control
What a person still decides.
The clinic draws the line, and the line is enforced by the platform rather than by a politely worded instruction.
- Anything clinical
- Symptoms, medication, whether a treatment is suitable after surgery. Scout says a clinician has to answer and hands the thread over, every time.
- Anything priced before assessment
- If a treatment cost depends on an examination, Scout quotes no figure. It books the assessment instead.
- Cancellations beyond policy
- Inside the window you publish, Scout can cancel and rebook. Outside it — late cancellations, fees, repeated no-shows — a person decides.
- A human path, always
- Any thread can be taken over mid-flow, and the patient can ask for a person at any point. The handoff carries the full history.
Source: WhatsApp Business Messaging Policy, updated 23 September 2026.
What you need
A channel, a diary Scout can read, and the two lists that define what it may say.
- A connected channel — WhatsApp, Instagram, Messenger, email or web chat.
- Opening hours, including the days each location is closed.
- The clinicians who can be booked, and what each of them is booked for.
- An availability endpoint over HTTPS: free slots out, a booking in.
- The questions you want asked before a slot is held — new or returning, which clinician, which location.
- The refusals you want enforced, starting with anything clinical.
Questions clinics actually ask.
Do you integrate with our practice-management system?
Not as a built-in connector. Scout calls an HTTPS tool you configure, so whatever your diary lives in, the integration is a small endpoint on your side that returns free slots and accepts a booking.
What stops it from answering a medical question?
A guardrail evaluated before the reply is sent. Clinical content is blocked at that layer rather than discouraged in a prompt, and the thread is handed to a named person with everything attached.
What happens if there is no availability?
The patient is told the diary is full for the days they asked about, and the request is assigned to the front desk. Scout does not offer a time the tool did not return.
Can a patient cancel or move a booking in the same thread?
Within the cancellation policy you configure, yes — it is the same tool call in reverse. Outside that policy it goes to a person, because the decision is yours, not the agent's.
Does the patient get a reminder?
A confirmed booking schedules a utility template ahead of the appointment, and the patient can reply to it to reschedule. That flow has its own page.
What resolution rate should I expect?
We do not publish one. Replay your own past booking conversations in an evaluation and read the score before you put this in front of patients.
Point it at your diary.
Book a demo and we will wire a test availability endpoint, then try to make the agent answer a clinical question in front of you.