The trigger finds customers whose comeback clock ran out, the filter removes the ones who shouldn't hear from you, and three messages do the rest.
Each customer is checked against the natural cycle for their service, not a blanket "hasn't booked in a year" date.
Rules run against your records before anything sends, so the list that survives is happy customers who simply drifted.
References the specific job. Proves it's your real company remembering a real visit.
Modest returning-customer credit or a seasonal deadline. A nudge, not a bribe.
Last touch. Nobody ever gets a fourth message.
Booked customers exit the sequence instantly and go back on their service clock. A "stop" reply removes them permanently, no exceptions.
The AI offers real openings from your calendar and books it. The job detail is already on the ticket.
Answered from your approved pricing and service list. Anything off-script gets flagged instead of guessed at.
Routes straight to a person with the sequence paused. That's a service recovery moment, not a scheduling one.
Each message sounds like your business, word for word, before the first send.
A $25 loyalty credit reads better than 40% off, which mostly tells people your prices have room in them.
After the first backlog send, customers cross their thresholds a few at a time, which books jobs year-round instead of spiking one week.