Automatic Waitlist for Cancelled Appointments, Run by an AI Worker

What the service is

The service is an AI worker connected to your booking calendar and your waiting list. When an appointment is cancelled, the worker finds the freed slot, works out which waitlisted clients could take it, and offers it to them one at a time. A client accepts by replying in the same chat. The worker then writes the booking into the calendar and records the change. Anyone on the list who never replies is passed over, and the next match receives the offer.

The point of the design is the reply. A message that only informs a client about an appointment cannot hold a slot for someone else. The system needs a conversation in which the client answers yes or no, and the booking is made only after that answer arrives.

Who it is for

The service suits any business that sells time with a person and loses slots to late changes. Clinics, dental and physiotherapy practices, salons, tutoring centres, legal and accounting offices, repair workshops and field-service firms all fit this pattern. A five-person practice and a group with several sites can both use it, provided the waitlist exists in some form and clients can be reached by phone number.

It is less useful where the waitlist is almost empty. If cancellations are rare, or if a short list can be handled by one person in ten minutes, the set-up effort may not be justified. The audit answers this question with your numbers before anything is built.

What is included

• A review of your current booking calendar, waiting list and cancellation patterns
• A written process map showing each trigger, matching rule, message and exception path
• A waitlist sheet or database with fields for day, time window, provider and visit type preferences
• Matching logic that checks each freed slot against those preferences before any message is sent
• Message templates for text and WhatsApp, written in the wording your clients will recognise
• An ordered offer sequence, with a set time window for each reply before the next client is contacted
• Automatic booking into the calendar when a client confirms, and release of the slot if the booking fails
• A daily log of offers, replies, bookings and skipped clients
• An exception queue for staff, covering unclear replies, clients who ask for other options, and bookings the calendar rejects
• Training notes for front-desk and operations staff, and a monthly review of the log

How it runs

1. Audit. We look at how cancellations arrive, how the waitlist is kept today, which calendar and messaging tools you use, and where your clients are. The output is a short written findings note.
2. Process map. We draw each step from cancellation to booking, including who approves what. You confirm the matching rules, such as whether a client who accepts any provider comes after a client who wants a named one.
3. Build. The waitlist, matching rules, message templates and calendar link are built and tested with dummy cancellations. Nothing is sent to real clients at this stage.
4. Run. The worker goes live for a supervised period. Staff check each day's log and handle every exception, and rules are adjusted where the log shows a pattern.
5. Support. After go-live, the worker is reviewed on a regular schedule. Changes to templates, providers or opening hours are made through a documented request.

How matching works in practice

Matching is the part most owners underestimate. A slot is not simply a time. It belongs to a provider, a visit type and a day, and clients on the list have their own limits. The table below shows how three common cases are handled.

Every rule in this table is set during the process map, not guessed by the system. Rules can be changed later, but each change should be written down and reviewed.

Why two-way messaging matters

A 2026 report by KevinMD on a multisite evaluation of automated waitlists described the results of an academic practice with 9,835 patients. No-show rates were 13.6 percent after a reminder call from front-desk staff, 17.3 percent after an automated reminder, and 23.1 percent with no reminder. The same report said automated waitlist software filled a median of 25 percent of the slots it offered for cancelled appointments. Those figures come from a different setting and a different tool. They show why a person still works the list when the software cannot close the gap, and why the reply step is worth designing carefully.

Messaging rules you must respect

On WhatsApp, Meta's Business Platform requires a user opt-in before a business sends message templates, and the opt-in must make clear who the business is and what it will send. Outside the customer service window, only approved template messages can be sent. In practice this means your clients need to have agreed to be contacted on WhatsApp, and the wording of the offer must be approved before it goes out. Text messages follow the rules of the carrier and the country. The audit records which channel each client has agreed to use.

What the client provides

• Access to the booking calendar, with permission for the worker to read open slots and create bookings
• The current waiting list, or the decision to start one, with each client's consent to be contacted
• Named providers, visit types and the rules for which appointments can be offered to whom
• Approved wording for messages, and sign-off on any template before it is submitted
• A named person who handles exceptions and reviews the log each week
• Clear opening hours and cancellation windows, so the worker does not offer a slot that has already passed

Typical timeline

The table gives a planning estimate for a single site with one calendar. Larger groups and multi-channel setups take longer, and the audit confirms the range for your business.

In total, a typical single-site setup reaches live operation in about six to eight weeks. Template approval by the messaging platform can add time, so it is started early.

What it does not cover

The service does not decide clinical priority, legal urgency or safety. A person must make those calls, and the worker is not set up to make them. It does not take payment, verify insurance or check eligibility unless you ask for those checks as separate steps. It cannot reach a client who has not agreed to be contacted, and it cannot fill a slot that no one on the list wants. It does not replace the front desk. It removes the repeated messaging and the first round of matching, so staff spend their time on the exceptions that need judgement.