Jane Voice AI with GoHighLevel: Booking, Cancellation and Rescheduling

A clinic voice AI receptionist needs more than a conversational script. It must collect the correct details, use those details in structured actions and interact with the clinic system without losing important information such as the practitioner, service, location or appointment time.

In the demonstrated workflow, a GoHighLevel voice AI agent uses ClinicSync Pro integrations to interact with Jane data. The test covers three common patient requests: booking an appointment, cancelling an existing appointment and moving an appointment to another time.

ClinicSync Pro is an independent product and is not affiliated with or endorsed by Jane or GoHighLevel.

For a new booking, the AI receptionist needs to identify the requested service, practitioner, location, date and time. If the caller does not provide everything at once, the conversation should gather the missing details before attempting the booking.

In the test, the caller requests a 30-minute massage with a named practitioner at a specific location and time. The workflow checks the requested availability, asks for the caller’s name, phone number and email address, confirms those details, and then completes the appointment action.

The resulting appointment is verified in Jane after the call. This verification is important during implementation because a spoken confirmation alone does not prove that the appointment was written to the clinic system successfully.

Patient matching is a critical part of appointment automation. If each call creates another patient record, clinic staff may be left with fragmented histories and duplicate contact details.

The tested booking reused an existing patient rather than creating another record for the same person. The demonstration then checks the patient search results to confirm that an additional duplicate was not created in that scenario.

Clinics should still test patient matching with realistic variations, including different phone number formats, spelling differences and updated email addresses. The video demonstrates one successful scenario rather than every possible patient-matching condition.

For cancellation requests, the AI receptionist first needs enough information to locate the patient and retrieve relevant appointments. When several upcoming appointments exist, presenting the available options helps the caller identify the correct one.

In the demonstration, the caller provides a first name and phone number. The workflow retrieves multiple upcoming appointments, the caller selects one by date, and the AI repeats the appointment details before processing the cancellation.

The cancelled status is then verified in Jane. This confirmation step reduces the risk of cancelling the wrong appointment when a patient has several similar bookings.

Rescheduling combines appointment lookup with the collection of a new date, time, practitioner and location preference. The workflow must identify the original appointment clearly and preserve any details the caller does not want to change.

During the test, the caller selects one of two upcoming appointments and requests a new date and time with the same practitioner and location. The AI confirms the requested change before completing the action.

The updated appointment is subsequently checked in Jane, where the new appointment time is visible and the previous appointment is shown as rescheduled.

The wording and tone of the receptionist can be customised, but the structured fields used by the prompt and custom actions must remain aligned. Required property names and data types are part of the workflow contract between the conversation and the action being called.

Changing a required field from the expected type, removing it or passing a different value format can prevent the action from receiving the information it needs. For example, appointment dates, identifiers and contact details should be passed in the format expected by the configured action.

Before editing a production prompt, document which parts control conversational style and which parts supply structured action data. Test every appointment action again after changing prompts or custom action fields.

The GoHighLevel account and Jane clinic settings should use aligned time zones. A mismatch can cause the caller’s requested time to be interpreted or displayed incorrectly.

The demonstration uses the same Pacific time zone in both systems. Clinics operating across multiple locations or time zones should pay particular attention to how relative phrases such as “tomorrow morning” are interpreted.

A practical test plan should include appointments near midnight, daylight-saving transitions and callers requesting dates without stating a time zone. Verify the resulting date and time directly in Jane.

Test booking with all details provided in one sentence, then repeat the test with details supplied gradually. Confirm that the workflow asks only for information that is still missing.

Test patients with one upcoming appointment and patients with several. The AI should identify the selected appointment and repeat its details before cancellation or rescheduling.

After every action, verify the result in Jane. Check the patient record, appointment status, practitioner, location, service, duration, date and time.

Repeat the process using an existing patient to check matching behaviour. Also test how the workflow handles incomplete, ambiguous or incorrectly spoken contact information.

Finally, retest booking, cancellation and rescheduling whenever prompt logic, custom action properties or account time-zone settings change.