AI Agent Use Cases
Views —

AI Agent Use Cases

Added today

10All 221

1–20 of 221

Work & ResearchWalkthrough

A bot that takes clinic bookings and checks for duplicates

The author shared how an n8n-based WhatsApp booking bot for a clinic handles duplicate webhooks and reduces double bookings.

n8n

How it works

  1. The webhook uses separate GET and POST nodes. POST returns 200 first, then verifies the signature and uses a spreadsheet to deduplicate by message ID.
  2. Dates and times are parsed with a rule-based Code node instead of AI. The bot checks Google Calendar availability before suggesting slots and, after a 'yes', checks again immediately before booking.
  3. After 3 misunderstood messages or a request for a human, the bot hands the conversation to staff through Telegram. A separate error workflow records failures in a spreadsheet and Telegram.

Results

According to the author, rechecking immediately before booking closed most of the double-booking window. Two requests arriving within the same second can still overlap.

Notes

It was tested on self-hosted n8n 2.43.7. Voice notes and images are not yet supported, so the bot asks for text. The workflow file was not published.

Source

Reddit Original post ↗

Example prompt · n8n

Build a WhatsApp booking bot for [industry] with n8n, ensuring that the same message is not processed twice and that two bookings are not made for the same time.

Inputs: Meta app for the WhatsApp Cloud API [app name], public HTTPS URL [URL], booking calendar [Google Calendar name], spreadsheet for state records [sheet name], Telegram bot for staff alerts [bot name], business hours and minimum booking interval [times], and language [language]. First check whether the installed n8n version supports webhook nodes, Code nodes, and Google Calendar availability queries. Explain the Meta message-template approval status and pricing.

Implementation: 1) Set up GET and POST webhooks on the same path; GET returns the challenge only when the verify token matches. 2) POST responds with 200 first, then verifies the signature using the app secret and discards mismatches. 3) Filter out delivery and read-status notifications; record message IDs in the spreadsheet and skip IDs already seen. Do not deduplicate by message text. 4) Read conversation state by phone number from the sheet, parse dates and times using rules, and suggest free slots. 5) When the customer confirms, recheck availability immediately before creating the event. 6) After three failures to understand or a request for a human, hand over to staff and stop the bot for that conversation. For conversations older than 24 hours, send only approved templates. 7) Record failures in the sheet and Telegram through a separate error workflow.

Deliverables: the workflow file, spreadsheet structure, and one paragraph of setup instructions. First use one test number to send the same message twice and verify it is processed only once. Attempt two bookings for the same time and confirm that only one is created before connecting the real number.