The AI receptionist setup playbook.

The gap between a bad AI receptionist and a great one is script quality and tuning discipline. The model is the same. Here’s how to set up Aaptly’s AI receptionist so it outperforms the human receptionist you used to pay $4,000/month for — in less than a week.

Day 1: Port or forward your number

The fastest path is call forwarding: point your existing number at the Aaptly AI receptionist. Takes 5 minutes with your carrier, caller ID stays yours, clients never notice. If you want full inbound control (voicemail, texting, etc.) port the number — that takes 2–5 business days but gives you everything in one place.

Day 1: Configure the script

The script is written in plain English in the Aaptly dashboard — no prompt engineering needed. Cover these sections:

• Greeting: “Hi, thanks for calling [Shop name]. I can help you book, reschedule, or answer questions.”
• Services: list every service you offer, with duration and price. Be specific — “Brazilian blowout, 2 hours, $220” beats “hair treatment, varies.”
• Providers: which providers do which services.
• Deposit policy: in plain English. “We collect a 25% deposit to hold the slot — applied to your total at checkout.”
• Cancellation policy: the window and the fee.
• FAQ: parking, directions, does-it-hurt, age requirements, whatever gets asked 3+ times a week.

Day 2: Test every scenario

Call the number yourself from every angle:

• A new client booking their first appointment.
• A returning client rescheduling.
• A client asking for a refund.
• A client asking about a service you don’t offer.
• A client speaking Spanish (or whichever second language your market needs).
• A cold call from a sales rep.

Note anything where the agent hesitates, misreads the intent, or gives a weird answer. Update the script in the dashboard. Repeat.

Day 3–5: Review transcripts

The agent logs every call as a searchable transcript. Spend 15 minutes a day for the first week scanning transcripts. You’re looking for:

• Missing FAQ topics — if 3+ callers ask the same thing and the agent can’t answer, add it.
• Price mismatches — the agent will read from your live config, but if pricing changed recently, double-check.
• Tone — does the agent sound like your shop? Script tweaks can shift it warmer, breezier, more professional.

Day 6: Turn on deposit collection

By now you trust the script. Enable mid-call deposit collection: the agent texts a Stripe payment link before hanging up, and the booking only locks once the deposit clears. This is the moment when the AI receptionist earns its keep — it starts recovering revenue that would’ve been lost to no-shows before the client even ends the call.

Day 7: Review the dashboard

At the end of week 1 you should see:

• 80%+ of calls resolved without human intervention.
• Booking conversion from calls around 65–75% (industry avg with a human: 40–50%).
• Zero missed calls — every caller got answered within 2 seconds.
• Recovered revenue tracking: deposits in, no-show fees charged.

The 20% of calls that weren’t fully resolved? They’re tagged in your unified inbox with transcript + intent classification, ready for you to reply in 30 seconds each.

Ongoing: the tuning loop

Review transcripts once a week after the first week. Add new services, update prices, expand the FAQ. Every update is instant — no engineer required, no deploy cycle. The agent gets smarter every week without you doing more than 10 minutes of work.