A tutoring center in Florida was losing an estimated $2,000 every month to a simple problem: nobody was available to book appointments between 7 PM and 11 AM, and parents looking for a tutor after hours were going straight to a competitor.
The Problem
Our client, a K-12 tutoring center in Florida, offered sessions across 50-plus subjects with around 15 tutors, each with completely different schedules, session lengths, location preferences, and availability rules.
Some tutors ran 45-minute sessions; others, 60-minute sessions, with start times on the hour or half-hour. Some worked online only, some in-person, some both. Classes had capacity limits; students could book single sessions, recurring weekly slots, or multi-session packages, and rescheduling required a minimum notice period to accommodate tutor travel.
The admin team handled all of this manually, creating a 16-hour window each day during which no bookings could be made. Parents who couldn’t get an immediate response moved on. The business ran on a first-come, first-served basis, and it was losing that race every night.
What We Built
We built a custom AI appointment system that handles inbound voice and text inquiries after hours, checks real availability across all tutors’ Google Calendars, enforces all scheduling rules, and automatically books confirmed appointments.
The scheduling logic was the real engineering challenge. The system needed to understand in real time which tutors were available for a given subject, day, format, session length, and start time, in a class that still had capacity. If a student already had a conflict or a class was full, the system had to surface the next viable option. Recurring bookings had to be validated across every future slot in the series, and rescheduling requests checked against the notice-period rule before being accepted.
Underneath the scheduling layer, we built a RAG knowledge system that covers the full subject library, so the AI can answer questions about subjects, tutors, and session formats accurately and consistently. The admin team retained full visibility and override control over every booking the AI made.
The Result
- The 16-hour after-hours booking gap closed completely, with the AI handling inquiries around the clock
- An estimated $1,000 to $2,000 per month in previously lost revenue recovered from missed leads
- Complex multi-tutor scheduling logic enforced consistently across session lengths, start times, location types, capacity limits, and rescheduling notice periods
- Accurate, consistent answers via a RAG knowledge base covering 50-plus subjects
- Full admin visibility and override control maintained throughout
The Honest Bit
This looks like a booking system on the surface, but the underlying scheduling logic is genuinely complex. A simple no-code tool handles a single calendar and a single session type. This system had to hold the full matrix of 15 tutors, each with their own rules, across dozens of overlapping subjects, with capacity limits, location constraints, recurring booking validation, and rescheduling logic all running simultaneously.
The hardest part wasn’t any single rule; it was building a system that could evaluate all of those conditions together in real time and return the right available slot, or the right explanation of why no slot existed, without failing silently or booking something invalid. Getting that logic airtight before going live took the most time of anything in the project.
Book your AI strategy call
Please schedule a AI consultation call. We’ll discuss your business needs and explore how AI can enhance your operations and drive growth.
