AI receptionist basics · 3 min
AI receptionist script templates for local businesses
A useful AI receptionist script tells the caller who they have reached, asks one clear question at a time, and sets an honest next step. These sample scripts are starting points, not ready-to-publish business policies. Replace every placeholder, remove actions your setup cannot perform, and have the owner approve the wording before use.
A short opening greeting
Template: 'Thanks for calling [Business Name]. I am the business's AI receptionist. I can help with general questions or take a request for the team. What can I help you with today?' Keep the opening brief enough for the customer to explain their reason for calling.
Only mention actions that are available in the actual setup. If the receptionist cannot book directly, do not open with a promise to schedule the customer. Add any business-approved notices required for the configured call experience; this template does not establish recording or messaging permission.
An appointment request without confirmed availability
Template: 'Which service are you interested in, and what day or time would suit you?' After collecting the answer: 'What name and contact method should the team use? I will pass along your preference. The team still needs to confirm availability.' Ask each question separately during the conversation.
If a real booking page is configured, an alternative is: 'You can check available appointments on our booking page.' Use the approved sharing method and do not say the appointment is confirmed merely because a link was offered. Confirmation must come from the booking provider or responsible staff.
An after-hours greeting
Template: 'The office is currently closed and reopens [Day and Time, Time Zone]. I can take a message for the team to review when they return. What would you like them to help with?' Replace the opening time with accurate information for weekends and holidays as well as ordinary weekdays.
If the business has a staffed on-call route, add only the owner-approved offer to try that route. Otherwise, avoid 'someone will call shortly' or other language that suggests overnight coverage. The script should distinguish taking a request from providing an immediate response.
A human transfer and its fallback
Template before transfer: 'I can try to connect you with the team. What would you like me to pass along?' If the configured transfer is unsuccessful: 'I could not reach someone. Would you like to leave a callback request? What is the best contact number and time for you?'
Use this wording only where the phone setup supports the transfer and fallback. Never announce that a person has answered before the connection succeeds. Test what the caller hears when the destination is busy, unanswered, or unavailable, and confirm that the callback reaches a responsible teammate.
An answer the receptionist does not know
Template: 'I do not have a confirmed answer to that question. I can pass it to the team so they can help.' Follow with the minimum details needed to route the question. For a price, policy exception, or technical decision, uncertainty should lead to a staff handoff rather than an invented answer.
This is a general fallback, not an emergency script. Have the business define separate instructions for safety concerns, sensitive complaints, and specialist questions. Review unanswered questions regularly and add approved responses where appropriate, without expanding the receptionist's authority beyond the team's intended limits.
Turn the templates into tested instructions
Before entering scripts in AeroDesk, replace the placeholders with the approved business name, hours, services, and contact process. Keep the receptionist instructions consistent with the saved booking URL and handoff destination. A script cannot enable a feature or connection that has not been configured.
Read each flow aloud and test interruptions, a request for a person, missing contact details, and a changed appointment preference. Check the resulting Inbox item as well as the spoken response. The script is ready only when the caller's stated next step matches the work the team actually receives.