AI after-hours receptionist

How the AI after-hours front desk works

Illustration with sample data.

Four AI helpers take a night of written sample calls for a made-up heating and air company. They answer from the business’s own answers, follow example rules the owner sets, and write a morning summary. A person approves the summary before the day team starts. The helpers never call, text or email anyone back.

Live phone calls are not built yet. These calls are written samples: no audio, nothing dialled.

Example rules for a heating and air company

• Emergency words like no heat, a leak or water coming in: page the on-call tech.
• An emergency with no callback number: put it first in the morning summary. No page, no promise of a callback.
• Booking requests: your rule. Book the open slot that fits, or hold it for you.
• No open slot fits, or the slot is taken: hold it for you.
• Questions: answer only from your own answers.
• A question your answers do not cover: hold it for you. Never guess.
• Recorded sales calls: drop them.
• No callback number: hold it, with the reason.
• Anything else: hold it for a person.

Who does what

• Maya, answers every call: Answers every call. Pages the on-call tech on emergency words, holds anything that needs a person, and hands the morning list over.
• Sara, screens callers: Screens callers. Drops recorded sales calls and holds a call with no callback number, with the reason.
• Paul, keeps the answers: Keeps the business’s own answers. Answers a known question word for word and holds anything else, never guessing.
• Joe, keeps the calendar: Keeps the calendar. Books or holds a request in the open slot that fits, by the owner’s rule, and holds it when nothing fits.

The service

The closest service today is a custom AI build: one task, built into the tools you already use.