Skip to content

Use case · Scheduling

Bookings, reminders and reschedules that keep your calendar right.

Customers ask for a time on chat, WhatsApp or the phone and get real openings from your calendar, not a form and a callback. The booking is written where your team already looks, reminders make moving a visit easier than skipping it, and the front desk keeps the final say.

At a glance

Connects to
  • Your calendar or booking system, via its API
  • Service rules table
  • Messaging channels
  • Phone line, optional
  • Customer records
  • Model provider on your account
Typical rollout
Week 1 → Week 5See the plan

The problem

Where the work gets stuck today

Booking by phone and message is slow back-and-forth. Finding a time takes several messages, after-hours calls go to voicemail, and the calendar is out of date because someone wrote the booking on a notepad first.

No-shows cost the slot and the revenue behind it. Reminders are either manual calls nobody has time for or a blunt text with no easy way to move the visit, so people who cannot make it simply do not come.

  • Requests sent in the evening are answered the next working day, when the slot has gone.
  • Two staff members offer the same opening to two different people.
  • Rescheduling means a phone call, so customers cancel or vanish instead.
  • The day list is only accurate once someone has rung round to confirm.

Task catalog

What the system takes on

  • Task 01

    Offer real openings

    Availability is read from your calendar with the service length, buffers, staff skills, opening hours and holidays applied, on chat, WhatsApp or the phone.

  • Task 02

    Hold the slot while booking

    The chosen time is held briefly while details are collected, so two people asking at once cannot both get it.

  • Task 03

    Collect only what the visit needs

    Name, contact and the reason for the visit, with returning customers matched by phone number or email.

  • Task 04

    Write and confirm

    The booking goes into the calendar of record, and the confirmation carries the address, what to bring and a reschedule link.

  • Task 05

    Remind with one-tap replies

    Reminders go out on the schedule you choose, with confirm, reschedule and cancel replies that update the calendar directly.

  • Task 06

    Move, cancel and refill

    Changes inside your policy window are handled automatically; a freed slot is offered to the waitlist before it goes back on general sale.

  • Task 07

    Give the front desk a day list

    Each morning the team sees who is confirmed, who moved and who has not replied, with the history on each booking.

Step trace

One run, step by step

Sample timeline of one booking: an after-hours WhatsApp request, three real openings, a held slot, the confirmation, a reminder, a one-tap reschedule to Thursday and the front-desk day list.
Sample data
  1. Mon 19:42 · WhatsApp

    A patient asks for a cleaning next week, after work

    The message arrives after the front desk has gone home.
  2. Mon 19:42 · Assistant

    Three real openings offered

    Read from the calendar with the visit length and clean-down buffer applied, inside the hygienist’s working hours.
  3. Mon 19:43 · Assistant

    Name, phone and reason for visit

    A returning patient is matched by phone number. Only the reason for the visit is asked, no clinical detail.
  4. Mon 19:43 · Calendar

    Booked, then confirmed back

    The event is written to the calendar of record; the confirmation carries the address, what to bring and a reschedule link.
  5. Tue 19:00 · Reminder

    Confirm, move or cancel in one tap

    Sent on the channel the patient used, in the wording the practice approved.
  6. Tue 19:06 · Patient

    Moves the visit to Thursday

    The same event is moved rather than a second one created. Wednesday reopens and is offered to the waitlist.
  7. Thu 08:30 · Front desk

    The day list shows it as confirmed and moved

    The change history sits on the booking, so nobody has to call to double-check.
Read the trace as text (Sample data)
  1. Mon 19:42PatientStepAsks on WhatsApp for a cleaning next week, after work.
  2. Mon 19:42AssistantStepOffers Tue, Wed and Thu at 17:30, read live from the calendar with buffers applied.
  3. Mon 19:43AssistantCheckHolds Wed 17:30 while it collects name, phone and reason for visit.
  4. Mon 19:43CalendarDoneRe-checks the slot, writes the event, sends the confirmation with a reschedule link.
  5. Tue 19:00ReminderStepReminder with confirm, reschedule and cancel replies.
  6. Tue 19:06PatientApprovalReschedules to Thu 17:30; the event is moved and Wednesday reopens for the waitlist.
  7. Thu 08:30Front deskDoneDay list shows the visit as confirmed and moved, with its history.

The calendar’s view

Step through the same booking on the calendar

The calendar stays the only source of availability. Every state the booking passes through is written back, so staff and the assistant always look at the same week.
  • Offered
  • Held
  • Booked
  • Reopened
  • Confirmed
calendar · next weekSample data
  1. Requested
  2. Offered
  3. Held
  4. Booked
  5. Reminded
  6. Moved
  7. Confirmed
Next week’s after-work slots for one hygienist (Sample data)
TimeMonTueWedThuFri
09:00Booked by someone elseOpenBooked by someone elseOpenOpen
12:30OpenBooked by someone elseOpenBooked by someone elseOpen
16:00OpenOpenBooked by someone elseOpenBooked by someone else
17:30Booked by someone elseOpenFilled from waitlistConfirmedBooked by someone else

Step 7 of 7

Confirmed, and the gap refilled

Thursday is confirmed on the day list. Wednesday went to the first person on the waitlist.

Systems involved

What it connects to

Your existing tools stay the system of record; the AI reads and writes through their APIs.
Your calendar or booking system, via its API
System of record for availability and bookings. The assistant reads open slots and writes, moves or cancels events; nothing is kept in a second calendar.
Service rules table
Visit lengths, buffers, which staff offer which service, opening hours, holidays and the cancellation window, editable by your team.
Messaging channels
Website chat, the WhatsApp Business Platform and SMS through Twilio for confirmations and reminders, only where the customer has agreed.
Phone line, optional
A voice agent on a Twilio number for callers who prefer to ring, using the same rules and the same calendar.
Customer records
Read to recognise returning customers by phone or email. Clinical notes are never read or written by the assistant.
Model provider on your account
OpenAI, Anthropic Claude or Google Gemini to understand requests such as “next Tuesday after five” and to word replies.

Guardrails

  • The calendar is the only source of availability. The assistant never promises a time it has not held, and it re-checks the slot before writing.
  • Service rules live in a table your team edits, not in a prompt, so a new holiday or a longer visit type takes effect at once.
  • Reminders go only on channels the person agreed to, in wording you approved, and an opt-out is honoured immediately.
  • For clinics the design is HIPAA-aware: the assistant asks only for the reason for the visit, never symptoms or history, and logs are masked.
  • Cancellations and reschedules inside your policy window are automatic; outside it, the request goes to staff with the conversation attached.
  • Every change is written back with the time and the channel, so staff can see who moved what and when.

Human checkpoint

The front desk keeps control of the calendar. Requests the rules cannot satisfy, such as a double-length visit, a named clinician who is fully booked, a late cancellation under your policy or anything clinical, go to a staff queue with the conversation attached. Staff block time or override a booking in the calendar exactly as they do today, and the assistant reads those changes on its next lookup.

What we plan for

Typical targets

Typical targets are planning ranges we use when scoping, each with its basis. They are not results from a client project; your own baseline decides the real number.
  • Typical target

    Every request, by design

    Booking requests answered outside opening hours

    Basis: The assistant answers on chat and WhatsApp at any hour and books into open slots; after-hours scenarios are part of the test script before launch.

  • Typical target

    None in the concurrency test

    Double bookings created by the assistant

    Basis: Slot holds plus a re-check before every write, tested with simultaneous booking attempts on the same slot before launch.

  • Typical target

    10–25% fewer

    No-shows on reminded appointments, against your own baseline

    Basis: Moving from no or manual reminders to automated reminders with one-tap rescheduling; reminders reach no-shows caused by forgetting or a clash. Measured against at least four weeks of your own no-show rate before launch.

Rollout

Week by week

Typical ranges for a first release; the scoping call fixes the plan for your systems.
  1. Week 1

    Service list, visit lengths, buffers, staff rules and cancellation window; calendar API access on an account in your name.

  2. Week 2

    Chat and WhatsApp booking flow against a test calendar; confirmation and reminder wording drafted for your approval.

  3. Week 3

    Reminders, reschedule and cancel replies, the waitlist, and the front-desk day list.

  4. Week 4

    Concurrency and edge-case tests (holidays, double-length visits, late cancellations), a staff walkthrough and a soft launch.

  5. Week 5

    Phone booking through a voice agent if it is in scope, and a review of the first fortnight of conversations.

FAQ

Frequently asked questions

Which calendars and booking systems can the assistant book into?

Any calendar or practice system with an API that exposes availability and lets us create, move and cancel bookings. If your system only allows reading, the assistant can collect the request and put it in a staff queue for one-click confirmation instead.

How does it stop two people booking the same slot?

The chosen slot is held for a short window while details are collected, and the calendar is checked again just before the booking is written. If someone else took it in the meantime, the customer is offered the nearest alternative instead of an error.

Can customers reschedule without calling us?

Yes. Every confirmation and reminder carries a reschedule option. The customer picks a new time from real openings, the existing booking is moved rather than duplicated, and the freed slot is offered to your waitlist.

When are reminders sent, and on which channel?

On the schedule you choose, commonly the day before and again a few hours before the visit, on the channel the customer booked through and agreed to. You approve the wording, and replies such as “confirm” or “move” update the calendar directly.

Can it handle several staff members or locations?

Yes. Each service is mapped to the staff who offer it and the locations where it runs, with their own hours. Customers can ask for a specific person, and if that person is fully booked the assistant offers their next opening or another qualified colleague, as your rules allow.

What happens when someone asks for a time we do not offer?

The assistant explains what is available and offers the nearest openings. If none suit, it can add the customer to the waitlist for that day or pass the request to staff, for example when a visit needs a double slot.

How much does the front desk have to change?

Very little. Staff keep using the same calendar, and the assistant reads their blocks and overrides on each lookup. The new parts are a small queue for requests the rules cannot handle and a morning day list with confirmations and changes.

Next step

Have a workflow in mind? Let’s scope it together.

Book a free 30-minute strategy call. Bring one process you want to automate and we will talk through scope, integrations and a sensible first step.

Free 30-minute call · no commitment

Send an RFP or security questionnaire