All Projects
PROJECT / 08

Pickleball Court Booking Automation

A connected n8n system where a Messenger AI agent books courts, holds slots, and confirms bookings only after payment is verified.

Built in n8n

A process worth improving.

Court bookings arrive through Messenger at all hours—someone has to check the calendar, quote the right price, hold the slot, and verify each payment by hand, and a slow reply can lose the slot to someone else.

A structured automation architecture.

Four linked n8n workflows: a Messenger AI agent answers with court knowledge and live pricing, checks availability against Google Calendar, and calls a booking-hold sub-workflow; separate sub-workflows record payment proof and verify each payment before the calendar slot is confirmed.

From trigger to reliable output.

Each stage has a defined responsibility, making the system easier to test, maintain, and improve.

01
Messenger message

Customer sends a booking request through Messenger.

Webhook
02
AI agent

Gemini agent with memory understands the request and replies.

AI
03
Rules & pricing

Court rules come from Google Docs, rates from Sheets.

Google Docs / Sheets
04
Availability check

Match the requested slot against court calendar events.

Google Calendar
05
Booking hold

Reserve the slot and save a pending booking record.

Calendar + Sheets
06
Payment proof

Record submitted payment proof against the booking.

Google Sheets
07
Verification

Verified payments update the booking and confirm the calendar slot.

Sheets + Calendar

Built as a system, not a shortcut.

The workflow is designed around predictable inputs, clear decisions, reviewable AI output, and recoverable errors.

  1. 01Trigger — receives and validates the initial input
  2. 02Data processing — structures information for downstream steps
  3. 03AI step — generates or classifies within defined constraints
  4. 04Conditions — route each result to the correct path
  5. 05API calls — exchange data with connected applications
  6. 06Output — delivers or stores a usable result
  7. 07Error handling — records failures and supports recovery
n8nFacebook Messenger APIGoogle GeminiGoogle CalendarGoogle SheetsGoogle Docs

The system is designed to give customers instant booking replies, hold court slots reliably, and confirm a booking only after its payment is verified.

Splitting one chatbot into focused sub-workflows keeps each booking stage testable on its own—and makes the money steps easy to audit.

Have a workflow like this?

Let’s Work Together