🧰 Before you start
- A voice-AI platform (e.g. Retell AI)
- Google Calendar or Cal.com
- An SMS-capable number
- Your Google review link
🪜 Step by step
- 1☐
Measure missed calls first
Check your phone records: how many calls went unanswered last month? That's your baseline.
- 2☐
Set up the receptionist
Follow the steps in the 'AI receptionist booked 200+ jobs' guide to create the agent and connect your calendar.
- 3☐
Define emergencies
Write exactly what counts as urgent and who gets alerted.
- 4☐
Add confirmation texts
Turn on SMS confirmations after booking and a reminder 24 hours before.
- 5☐
Add review requests
After each job, send a text with your Google review link.
- 6☐
Forward unanswered calls
Set your phone to forward to the AI after 2 rings or when busy.
- 7☐
Compare after 90 days
Check calls answered, bookings and reviews against your baseline.
💬 Copy-paste prompts
Tap Copy prompt, paste it into your AI (Grok, ChatGPT, Claude or Gemini), then replace the [yellow bits] with your own details.
You are [name], the friendly receptionist for [business] in [area]. Business hours: [hours]. On every call: 1. Greet the caller and ask how you can help. 2. Find out: their name, phone number, what they need, where they are (service area: [areas]) and how urgent it is. 3. If it's an emergency ([define, e.g. burst pipe]), say someone will call back within [15] minutes and send an alert to [number]. 4. Otherwise check the calendar and offer two available times. Repeat the time back and get a clear 'yes' before booking. 5. Never quote prices except: [standard prices]. For anything else, say the team will confirm. 6. Tell callers they are speaking with an AI assistant if they ask, and offer to take a message for a person. Keep answers short and friendly, in British English.
Write 3 short, friendly SMS messages (under 160 characters) asking a customer to leave a Google review after a [service] visit from [business]. Include [review link]. British English.
⚠️ Tips & common mistakes
- Ask customers for consent before sending marketing texts.
- Keep a human on call for true emergencies.
- Read transcripts to spot questions the AI struggles with.