Launch an AI Receptionist Business

If you must monitor every turn → build a custom GPT for chat; If steps are pure logic → build a workflow; If fixed order but needs AI calls → build an AI workflow; If you need autonomy + flexible dialog + tools → build an AI agent (receptionist).

System: You are an AI receptionist for [CLINIC NAME]. Goals: 1) Greet warmly, 2) Identify reason for calling, 3) Qualify urgency (pain, bleeding, trauma), 4) Offer next available appointment per [BOOKING RULES], 5) Capture: full name, phone, email, insurance (Y/N), preferred time, 6) If emergency → transfer to [HUMAN_NUMBER], 7) Summarize outcome for CRM, 8) End call politely.
Tools: call_transfer(number), end_call().
Behavior: Be concise, professional, no medical advice. If unsure, ask one clarifying question. Use today’s date/time: {{current_time_America/Toronto}}. If voicemail detected, end.
Handoff triggers: emergency keywords, billing disputes, cancellations <24h, hostile caller.
Closure: Confirm details back, provide appointment time, directions if asked, then end_call().

System: You are an AI receptionist for [SPA NAME]. Book consultations and treatments following [MENU/RULES]. Capture: name, phone, email, interest (Botox/Facial/Laser), budget range, best day/time. Offer packages if asked; never upsell aggressively. Urgent reactions → call_transfer([HUMAN_NUMBER]). Always end with clear next steps and end_call().

System: You are an AI receptionist for [COMPANY]. Qualify jobs (service type, location ZIP, urgency, access). Provide earliest availability within [SERVICE_AREA & SCHEDULE]. Transfer gas leaks/floods/safety to [HUMAN_NUMBER]. Confirm estimate windows and end_call().

System add‑on: Answer only from [KNOWLEDGE SOURCES]. If source is missing, say you’ll pass this to a human and create a handoff note. Never invent facts. If repeated uncertainty → call_transfer([HUMAN_NUMBER]).

Responsiveness: 0.85–0.95; Interruption sensitivity: 0.8–0.85; Back‑channel: on (low); Speech normalization: 0.3; Ring: 30s; End on silence: 30–60s; Max duration: 12–20 min; Voicemail detection: on if you want hang‑up on voicemail; off if you want to leave no voicemail.

Sheet: Receptionist CRM
Columns: timestamp, call_direction, caller_phone, caller_name, email, intent, urgency, service_interest, notes, scheduled_time, outcome, handoff_to, agent_id, recording_url, transcript_url, followup_status

1) In Google Cloud Console: create project → enable APIs (Sheets, Gmail).
2) Configure OAuth consent (External), app name, support email.
3) Create OAuth client (Web), add your automation redirect URI (n8n/Make/Zapier).
4) Copy client ID/secret into your platform’s Google credentials.
5) Approve scopes on first run (Advanced → proceed if unverified warning appears).

1) Create API key in OpenAI account and add a payment method.
2) Store in your automation platform’s OpenAI node/HTTP header.
3) Prefer specific models (e.g., GPT‑4.1) for consistent quality.

{
  "from": "[YOUR_AGENT_NUMBER]",
  "to": "+1{{$json.caller_phone}}",
  "callType": "phone",
  "agentId": "[AGENT_ID]",
  "vars": {
    "first_name": "{{$json.caller_name}}",
    "service_interest": "{{$json.service_interest}}",
    "notes": "{{$json.notes}}"
  }
}

Subject: [Personalized hook for [Name]]

Hi {{caller_name}}, I’m [AGENT NAME], the AI receptionist for [BUSINESS]. I’ll give you a quick call in ~10 minutes about [SERVICE]. If now’s bad, reply with a better time.

Subject: New call — {{caller_name}} ({{intent}})
Booked: {{scheduled_time || 'No'}}
Urgency: {{urgency}}
Service: {{service_interest}}
Notes: {{notes}}
Recording: {{recording_url}}
Transcript: {{transcript_url}}

1) Add contact form to your site builder.
2) Point the action to your production webhook URL.
3) Map fields to your Sheet schema.
4) On submit → append row → trigger pre‑call SMS → queue outbound call.

For {{client_name}}: Calls answered {{n}}, Booked {{booked_n}}, Handoffs {{handoff_n}}, Voicemail {{vm_n}}. Top intents: {{top_intents}}. Next actions: {{owners_and_dates}}.

Subject: [Personalized hook for [Name]]

Hey [NAME], noticed you’re likely missing after‑hours calls. I build AI receptionists that answer, qualify, and book appointments automatically. Want a quick demo tailored to [BUSINESS]? 5–10 mins.

Subject: Hey {{name}} I think I have something for you regarding {{unique thing about them or their company}}
Hi {{firstName}},

I know you’re {{icebreaker}} so I figured it made sense to chat.

The TLDDR, I think I can add 5K a month to your {{elevatorPitch}} with a few automated systems.

I just did this for a very similar industry company and we had 28,350 in a few months. They do related things. So I’m very confident I can duplicate this at minimum. Would be 100% risk-free. I guarantee at least 20 appointments booked or you wouldn’t have to pay.

P.S. {{postscript}}

Who do you know that’s looking for an AI‑powered receptionist that helps them capture missed calls and book more appointments?

Title: AI Receptionist to stop missed calls at [BUSINESS]
Intro: 2–3 lines on missed calls ROI for their niche.
Mini‑demo: 60–90s Loom: show booking flow + Sheet log + summary email.
Offer: 14‑day pilot, then monthly retainer with clear deliverables.
CTA: Book a quick fit call (link).

Create a 5–10 slide pitch deck showing current missed‑call problems on [BUSINESS] and how an AI receptionist could increase bookings by 15–30%. Include pain points by slide and solutions toward the end.

Starter ($500/mo): AI receptionist + missed‑call text back + basic site/contact form hookup.
Growth ($1,500–$2,000/mo): Receptionist + lead gen campaigns + reporting.
All‑in ($2,500–$3,000/mo): Receptionist + ads + SEO + advanced automations.
Value‑based: ~25% of savings captured (adjust by risk tolerance).

Thanks for partnering with us. Cadence: weekly check‑ins for 4 weeks, then biweekly. Today we’ll align on goals, connect accounts, and go over security. We’ll share a go‑live date, and how to reach us if anything urgent happens. Q&A at the end.

OpenAI: user‑facing chat & personality; Claude: complex coding/agents & structured JSON; Gemini: long context & cost‑effective research. Mix intentionally.

{
  "caller_name": "...",
  "caller_phone": "...",
  "intent": "...",
  "urgency": "low|medium|high",
  "service_interest": "...",
  "scheduled_time": "ISO8601|null",
  "handoff_to": "name|null",
  "notes": "..."
}

Subject: [Personalized hook for [Name]]

Hey, I noticed you might be missing leads after business hours. I help businesses automate lead follow‑up with an AI receptionist so you can book more appointments automatically. Want to see it on your site/number?