How a driving instructor cut manual texting by 70%: the full build

Our favourite case study, in detail. One instructor, fully booked, spending evenings answering texts between lessons: and losing enquiries he could not answer fast enough.

The problem, measured

Before building anything we counted: roughly 40 inbound texts a day, most asking about availability. Every reply was manual. Double bookings happened about twice a month, each one an apology and a lost hour.

The build

• SMS assistant: answers enquiries in his voice, connected to his real calendar through Twilio and a secure AI workflow.

• Live availability: the assistant offers only genuinely free slots, so double bookings became structurally impossible.

• Reminders: automatic texts the evening before and an hour before each lesson.

• Human handoff: anything unusual (complaints, reschedules of test-day lessons) gets flagged to him instead of answered automatically.

The results

Seventy percent of messages now handled without him touching the phone. Zero double bookings since launch. Enquiries answered in seconds instead of hours: which, for a local service business, is the difference between winning and losing the job.

What it did not do

It did not replace him, change how he teaches, or touch anything requiring judgement. That is the point: automate the coordination, keep the human for the humans.