klbr/n8nworkflows.xyz
1
0
Fork 0
mirror of https://github.com/khoaliber/n8nworkflows.xyz.git synced 2026-04-28 00:29:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

creation

Browse Source
This commit is contained in:
nusquama
2025-11-13 14:51:05 +01:00
parent ad4c93d450
commit 7b4d737873
1 changed files with 415 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
workflows/Automate WhatsApp Booking System with GPT-4 Assistant, Cal.com and SMS Reminders-4949/readme-4949.md
+415
View File
@@ -0,0 +1,415 @@
Automate WhatsApp Booking System with GPT-4 Assistant, Cal.com and SMS Reminders
https://n8nworkflows.xyz/workflows/automate-whatsapp-booking-system-with-gpt-4-assistant--cal-com-and-sms-reminders-4949
# Automate WhatsApp Booking System with GPT-4 Assistant, Cal.com and SMS Reminders
### 1. Workflow Overview
This n8n workflow automates the booking process via WhatsApp using an AI assistant powered by GPT-4, integrates with Cal.com for appointment scheduling, and sends SMS reminders before appointments. It targets service providers who want to handle client bookings conversationally and automatically while managing follow-ups through reminders.
The workflow is logically divided into three functional blocks:
- **1.1 User Interaction and Qualification:** Receives WhatsApp inputs, interprets user intent, and guides users through service selection and personal information collection via an AI assistant.
- **1.2 Appointment Booking:** Automatically books the appointment with Cal.com using collected data, updates Google Sheets with booking details, and confirms the booking with the user.
- **1.3 SMS Reminders:** Scheduled hourly trigger that reads upcoming confirmed appointments from Google Sheets, filters those occurring within the next two hours, sends SMS reminders, and marks reminders as sent.
---
### 2. Block-by-Block Analysis
#### 2.1 User Interaction and Qualification
**Overview:**
This block handles incoming WhatsApp messages, determines if the user is chatting or confirming a booking, runs the AI assistant to qualify the user (collecting name, email, and service), and stores prospect data.
**Nodes Involved:**
- Webhook Trigger (WhatsApp Input)
- Switch: Confirm vs Chat Flow
- AI Booking Assistant (Dr Firas)
- LLM: GPT-4o Chat Model
- AI Conversation Memory
- Create New Prospect in Google Sheet
- Update Prospect with Booking Details
- Fetch Available Time Slots
- Send Response to WhatsApp
**Node Details:**
- **Webhook Trigger (WhatsApp Input)**
- Type: Webhook Trigger
- Role: Entry point to receive WhatsApp POST requests containing user messages and metadata.
- Config: HTTP POST at specific path, responds via response node.
- Inputs: External WhatsApp API integration.
- Outputs: Passes JSON with user input and contact info downstream.
- Potential failures: Invalid webhook calls, missing data, connectivity issues.
- **Switch: Confirm vs Chat Flow**
- Type: Switch node
- Role: Routes flow based on presence of "confirm" keyword in user input.
- Config: Two outputs — "Discussion" if user input does not contain "confirm", "Confirm booking" if it does.
- Inputs: Webhook Trigger node output.
- Outputs: Routes to AI assistant for chat or to prospect retrieval for confirmation.
- Edge cases: Case sensitivity or unexpected inputs; relies on exact string matching.
- **AI Booking Assistant (Dr Firas)**
- Type: Langchain Agent node (AI assistant)
- Role: Processes user input with GPT-4o model, follows a detailed scripted conversation flow to qualify user and suggest appointments.
- Config: Custom system prompt instructing assistant on services, tone, conversation flow, and constraints. It collects name, email, service selection, offers time slots, and requests confirmation.
- Inputs: User message text.
- Outputs: AI-generated replies to be sent back to WhatsApp.
- Uses: LLM: GPT-4o Chat Model for language generation, AI Conversation Memory for session context, Fetch Available Time Slots tool to get appointment times, and Google Sheets nodes for data storage.
- Edge cases: AI misunderstanding, session loss, API rate limits, incomplete user data.
- **LLM: GPT-4o Chat Model**
- Type: Langchain OpenAI LLM node
- Role: Provides GPT-4o model responses for the AI assistant.
- Config: Model set to "gpt-4o".
- Credentials: OpenAI API key.
- Inputs: Text prompt from AI Booking Assistant.
- Outputs: AI text completion.
- Edge cases: API limits, timeout, malformed prompts.
- **AI Conversation Memory**
- Type: Langchain memory buffer
- Role: Maintains conversation state per user to enable contextual replies.
- Config: Uses WhatsApp contactId as session key, window length of 50 messages.
- Inputs/Outputs: Connects to AI Booking Assistant.
- Edge cases: Memory overflow, session key mismatch.
- **Create New Prospect in Google Sheet**
- Type: Google Sheets Tool node
- Role: Stores prospect data (name, email, phone, service, event ID, conversation summary) after collection.
- Config: Append or update mode matching on contact ID, targeting specific sheet and document.
- Inputs: Data from AI assistant.
- Credentials: Google Sheets OAuth2.
- Potential errors: Google API auth failure, sheet access issues.
- **Update Prospect with Booking Details**
- Type: Google Sheets Tool node
- Role: Updates existing prospect record with booking date/time and event ID after user selects a time slot.
- Config: Update mode with matching on contact ID.
- Inputs: AI assistant output.
- Credentials: Google Sheets OAuth2.
- **Fetch Available Time Slots**
- Type: HTTP Request tool node (Langchain tool)
- Role: Queries Cal.com API to retrieve available appointment slots for selected service.
- Config: GET request with parameters eventTypeId, startTime (now), endTime (+48h), max 5 slots, uses Paris timezone in ISO 8601 format.
- Headers: Authorization Bearer token, Content-Type JSON.
- Inputs: Event ID and time params from AI assistant.
- Outputs: Available slots data for AI to display.
- Potential issues: API rate limits, invalid event ID, time format errors.
- **Send Response to WhatsApp**
- Type: Respond to Webhook node
- Role: Sends AI assistant's replies back to the WhatsApp user as response to webhook call.
- Inputs: AI assistant output.
- Outputs: HTTP response containing message(s).
---
#### 2.2 Appointment Booking
**Overview:**
This block normalizes the user-selected booking time, sends booking data to Cal.com, checks booking success, updates Google Sheets with confirmed bookings, and responds back to the user.
**Nodes Involved:**
- Retrieve Prospect Details
- Normalize Booking Time (UTC Format)
- Send Booking
- Format Date for Readability (Display)
- Check Booking Status (Success or Error)
- Webhook Response: Booking Confirmed
- Webhook Response: Booking Failed
- Mark Booking as Confirmed in Sheet
**Node Details:**
- **Retrieve Prospect Details**
- Type: Google Sheets node
- Role: Retrieves prospect’s stored data by contact ID to use for booking creation.
- Config: Sheet filter on contact ID from WhatsApp input.
- Credentials: Google Sheets OAuth2.
- **Normalize Booking Time (UTC Format)**
- Type: Code node (JavaScript)
- Role: Converts the booking date/time string (possibly in Paris timezone) into ISO 8601 UTC format with 'Z' suffix for Cal.com API.
- Inputs: "Date du rendez-vous" from prospect details.
- Outputs: JSON with original and normalized date strings.
- Edge cases: Invalid or malformed date input.
- **Send Booking**
- Type: HTTP Request node
- Role: Posts booking data to Cal.com API to create an appointment.
- Config: POST with JSON body including eventTypeId, normalized start time, attendee info (name, email, timezone Europe/Paris), and booking field responses (summary).
- Headers: Authorization Bearer token, content-type JSON, Cal API version.
- Inputs: Normalized date, prospect details.
- Outputs: Booking API response including status.
- Failures: API errors, invalid data, network timeout.
- **Format Date for Readability (Display)**
- Type: Code node (JavaScript)
- Role: Formats the appointment date/time into a human-readable string (e.g., "June 4, 9am") for user confirmation messages.
- Inputs: Raw booking date from prospect details.
- Outputs: JSON with original and formatted date strings.
- Edge cases: Invalid date inputs.
- **Check Booking Status (Success or Error)**
- Type: Switch node
- Role: Routes flow based on Cal.com booking response status (success or error).
- Inputs: Send Booking node output.
- Outputs: To success or failure webhook responses.
- **Webhook Response: Booking Confirmed**
- Type: Respond to Webhook node
- Role: Sends a JSON confirmation message to the user acknowledging successful booking.
- **Webhook Response: Booking Failed**
- Type: Respond to Webhook node
- Role: Sends JSON error message to the user if booking failed (e.g., slot no longer available).
- **Mark Booking as Confirmed in Sheet**
- Type: Google Sheets node
- Role: Updates the prospect’s row marking the booking as confirmed (“confirm”) for record keeping.
- Inputs: Contact ID from Retrieve Prospect Details.
- Credentials: Google Sheets OAuth2.
---
#### 2.3 SMS Reminders
**Overview:**
This block runs on a schedule every hour, reads upcoming confirmed appointments from the sheet, filters those within the next 2 hours, sends SMS reminders to clients, and marks reminders as sent.
**Nodes Involved:**
- Trigger: Check Appointments Every Hour
- Read Upcoming Appointments from Sheet
- Filter Appointments Within Next 2 Hours
- Send SMS Reminder
- Mark SMS as Sent in Sheet
**Node Details:**
- **Trigger: Check Appointments Every Hour**
- Type: Schedule Trigger node
- Role: Initiates the reminder flow every hour automatically.
- **Read Upcoming Appointments from Sheet**
- Type: Google Sheets node
- Role: Reads all appointment records to find candidates for reminders.
- Credentials: Google Sheets OAuth2.
- **Filter Appointments Within Next 2 Hours**
- Type: Code node (JavaScript)
- Role: Filters appointments that are:
- Confirmed ("Réservé" = "confirm")
- Have no reminder sent yet ("Rappel SMS envoyé" ≠ "envoyer")
- Scheduled within the next 2 hours according to Paris timezone.
- Also formats appointment time for logging and message.
- Outputs filtered list of appointments needing reminders.
- Edge cases: Invalid date formats, timezone miscalculations.
- **Send SMS Reminder**
- Type: sms77 node
- Role: Sends SMS to the prospect’s phone number with a reminder message including appointment details.
- Credentials: sms77 API key.
- Inputs: Phone number and personalized message from filtered appointments.
- Failures: SMS API errors, invalid phone numbers.
- **Mark SMS as Sent in Sheet**
- Type: Google Sheets node
- Role: Updates prospect’s record marking that SMS reminder has been sent ("envoyer") to prevent duplicates.
- Credentials: Google Sheets OAuth2.
---
### 3. Summary Table
| Node Name | Node Type | Functional Role | Input Node(s) | Output Node(s) | Sticky Note |
|----------------------------------|----------------------------------|----------------------------------------------|---------------------------------------|-----------------------------------------|--------------------------------------------------------------------------------------------------|
| Webhook Trigger (WhatsApp Input) | Webhook Trigger | Entry point for WhatsApp messages | - | Switch: Confirm vs Chat Flow | # 🟫 STEP 1 — Qualify User and Suggest Appointment |
| Switch: Confirm vs Chat Flow | Switch | Routes chat or booking confirmation flows | Webhook Trigger | AI Booking Assistant, Retrieve Prospect Details | |
| AI Booking Assistant (Dr Firas) | Langchain Agent | Conversational AI to qualify user & suggest appointments | Switch: Confirm vs Chat Flow | Send Response to WhatsApp | |
| LLM: GPT-4o Chat Model | Langchain LLM | Language model for AI assistant | AI Booking Assistant | AI Booking Assistant | |
| AI Conversation Memory | Langchain Memory | Maintains conversational context | AI Booking Assistant | AI Booking Assistant | |
| Create New Prospect in Google Sheet | Google Sheets Tool | Stores prospect data after qualification | AI Booking Assistant | AI Booking Assistant | |
| Update Prospect with Booking Details | Google Sheets Tool | Updates prospect with selected booking details | AI Booking Assistant | AI Booking Assistant | |
| Fetch Available Time Slots | Langchain HTTP Request Tool | Retrieves available appointment slots from Cal.com | AI Booking Assistant | AI Booking Assistant | |
| Send Response to WhatsApp | Respond to Webhook | Sends AI assistant's reply to WhatsApp user | AI Booking Assistant | - | |
| Retrieve Prospect Details | Google Sheets | Retrieves prospect info for booking | Switch: Confirm vs Chat Flow | Normalize Booking Time (UTC Format) | |
| Normalize Booking Time (UTC Format) | Code | Converts booking time to UTC ISO string | Retrieve Prospect Details | Send Booking | |
| Send Booking | HTTP Request | Sends booking request to Cal.com API | Normalize Booking Time | Format Date for Readability (Display) | |
| Format Date for Readability (Display) | Code | Formats booking date/time to readable string | Send Booking | Check Booking Status | |
| Check Booking Status (Success or Error) | Switch | Routes flow on booking success or failure | Format Date for Readability | Webhook Response: Booking Confirmed, Webhook Response: Booking Failed | |
| Webhook Response: Booking Confirmed | Respond to Webhook | Sends confirmation message to user | Check Booking Status | Mark Booking as Confirmed in Sheet | |
| Webhook Response: Booking Failed | Respond to Webhook | Sends failure message to user | Check Booking Status | - | |
| Mark Booking as Confirmed in Sheet | Google Sheets | Marks booking as confirmed in sheet | Webhook Response: Booking Confirmed | - | |
| Trigger: Check Appointments Every Hour | Schedule Trigger | Hourly trigger to check upcoming appointments | - | Read Upcoming Appointments from Sheet | # 🟩 STEP 3 — Send SMS Reminder Before Appointment |
| Read Upcoming Appointments from Sheet | Google Sheets | Reads all appointments from sheet | Trigger: Check Appointments Every Hour | Filter Appointments Within Next 2 Hours | |
| Filter Appointments Within Next 2 Hours | Code | Filters appointments within next 2 hours needing reminder | Read Upcoming Appointments from Sheet | Send SMS Reminder | |
| Send SMS Reminder | sms77 | Sends SMS reminders to clients | Filter Appointments Within Next 2 Hours | Mark SMS as Sent in Sheet | |
| Mark SMS as Sent in Sheet | Google Sheets | Marks SMS reminder as sent in sheet | Send SMS Reminder | - | |
| Sticky Note | Sticky Note | Visual step divider/label (Step 1) | - | - | # 🟫 STEP 1 — Qualify User and Suggest Appointment |
| Sticky Note3 | Sticky Note | Visual step divider/label (Step 2) | - | - | # 🟥 STEP 2 — Book Appointment Automatically |
| Sticky Note (Step 3) | Sticky Note | Visual step divider/label (Step 3) | - | - | # 🟩 STEP 3 — Send SMS Reminder Before Appointment |
---
### 4. Reproducing the Workflow from Scratch
1. **Create Webhook Trigger (WhatsApp Input):**
- Type: Webhook Trigger
- Method: POST
- Path: Unique identifier (e.g., "6438cd95-74cb-4f40-a1a5-853706fe96f6")
- Response Mode: Respond via response node
- Purpose: Receive WhatsApp messages with JSON body containing user input and contact info.
2. **Add Switch Node (Confirm vs Chat Flow):**
- Type: Switch
- Conditions:
- Output "Discussion": If userInput does NOT contain "confirm" (case-sensitive).
- Output "Confirm booking": If userInput contains "confirm".
- Connect input from Webhook Trigger.
3. **Set Up AI Booking Assistant (Dr Firas):**
- Type: Langchain Agent node
- Configuration:
- Text input: `={{ $json.body.userInput }}`
- System message: Embed detailed scripted conversation in French including identity, services, tone, flow, constraints, and date.
- Connect input from "Discussion" output of Switch node.
4. **Configure LLM: GPT-4o Chat Model:**
- Type: Langchain LLM node
- Model: "gpt-4o"
- Credentials: Link to OpenAI API key
- Connect input from AI Booking Assistant node's LLM input.
5. **Add AI Conversation Memory:**
- Type: Langchain Memory Buffer
- Session Key: `={{ $json.body.contactId }}`
- Context Window Length: 50
- Connect input/output to AI Booking Assistant node.
6. **Add Fetch Available Time Slots Node:**
- Type: Langchain HTTP Request Tool
- URL: `https://api.cal.com/v2/slots/available`
- Method: GET with query parameters: eventTypeId, startTime, endTime
- Headers: Authorization Bearer token (Cal.com live token), Content-Type application/json
- Connect as AI Tool input to AI Booking Assistant node.
7. **Create New Prospect in Google Sheets:**
- Type: Google Sheets Tool (append or update)
- Document ID and Sheet Name: Your Google Sheet for appointments
- Matching Column: "ID du contact"
- Columns to capture: Nom, Courriel, Numéro de téléphone, Résumé, ID du contact
- Credentials: Google Sheets OAuth2
- Connect as AI Tool output to AI Booking Assistant node.
8. **Update Prospect with Booking Details:**
- Type: Google Sheets Tool (update)
- Columns: ID du contact, Date du rendez-vous, ID de l’événement, Nom de l’événement
- Matching Column: "ID du contact"
- Credentials: Google Sheets OAuth2
- Connect as AI Tool output to AI Booking Assistant node.
9. **Send Response to WhatsApp Node:**
- Type: Respond to Webhook
- Respond with: All incoming items
- Connect output from AI Booking Assistant main output.
10. **Retrieve Prospect Details:**
- Type: Google Sheets node
- Filter: "ID du contact" equals `={{ $json.body.contactId }}` from webhook
- Credentials: Google Sheets OAuth2
- Connect from "Confirm booking" output of Switch node.
11. **Normalize Booking Time (UTC Format) Code Node:**
- JavaScript to convert "Date du rendez-vous" to ISO 8601 UTC string with 'Z' suffix.
- Connect output from Retrieve Prospect Details.
12. **Send Booking HTTP Request Node:**
- URL: `https://api.cal.com/v2/bookings`
- Method: POST
- JSON Body: eventTypeId, start (normalized time), attendee (name, email, timezone Europe/Paris), bookingFieldsResponses (title)
- Headers: Authorization Bearer token, Content-Type application/json, cal-api-version
- Connect output from Normalize Booking Time node.
13. **Format Date for Readability (Display) Code Node:**
- JavaScript formatting date string to readable format (e.g., "June 4, 9am") using timezone Europe/Berlin (close to Paris).
- Connect output from Send Booking node.
14. **Check Booking Status Switch Node:**
- Condition:
- Output "succeeded" if `status` equals "success" in Send Booking response.
- Output "Failed" if `status` equals "error".
- Connect output from Format Date node.
15. **Webhook Response: Booking Confirmed Node:**
- Type: Respond to Webhook
- Respond with JSON confirming booking success message.
- Connect from "succeeded" output of Switch.
16. **Webhook Response: Booking Failed Node:**
- Type: Respond to Webhook
- Respond with JSON error message indicating slot unavailable.
- Connect from "Failed" output of Switch.
17. **Mark Booking as Confirmed in Sheet:**
- Type: Google Sheets node (update)
- Set "Réservé" column to "confirm" for matching contact ID.
- Credentials: Google Sheets OAuth2
- Connect from Webhook Response: Booking Confirmed node.
18. **Trigger: Check Appointments Every Hour:**
- Type: Schedule Trigger
- Interval: Every 1 hour
19. **Read Upcoming Appointments from Sheet:**
- Type: Google Sheets node
- Read entire sheet containing appointments.
- Credentials: Google Sheets OAuth2
- Connect from schedule trigger.
20. **Filter Appointments Within Next 2 Hours Code Node:**
- JavaScript filtering confirmed appointments ("Réservé" = "confirm"), no reminder sent yet ("Rappel SMS envoyé" ≠ "envoyer"), within the next 2 hours in Paris timezone.
- Also formats appointment display time for message.
- Connect from Read Upcoming Appointments node.
21. **Send SMS Reminder Node:**
- Type: sms77 node
- To: Prospect phone number from filtered items
- Message: Personalized reminder with name and event name, note appointment is within 2 hours.
- Credentials: sms77 API key
- Connect from filtered appointments node.
22. **Mark SMS as Sent in Sheet:**
- Type: Google Sheets node (update)
- Sets "Rappel SMS envoyé" field to "envoyer" for matching contact ID.
- Credentials: Google Sheets OAuth2
- Connect from Send SMS Reminder node.
23. **Add Sticky Notes:**
- Create three sticky notes for visual separation of steps in workflow:
- Step 1: Qualify User and Suggest Appointment
- Step 2: Book Appointment Automatically
- Step 3: Send SMS Reminder Before Appointment
---
### 5. General Notes & Resources
| Note Content | Context or Link |
|-----------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------|
| Workflow handles 24/7 WhatsApp bookings with AI assistant in French and English, using GPT-4o. | Core functionality and language context. |
| Cal.com API used for scheduling availability and bookings with ISO 8601 dates in Paris timezone. | https://cal.com/docs/api/ |
| SMS reminders sent via sms77 API every hour for upcoming appointments within 2 hours. | https://www.sms77.io/en/docs/ |
| Google Sheets used as CRM for prospect and booking data with OAuth2 credentials. | Google Sheets API docs: https://developers.google.com/sheets/api |
| AI assistant system prompt includes detailed conversational flow and constraints for user guidance. | Custom prompt embedded inside AI Booking Assistant node. |
| Important to manage API rate limits and error handling especially for OpenAI, Cal.com, and sms77. | Consider retry logic or alerts for failures in production. |
---
**Disclaimer:**
The provided text is extracted exclusively from an automated workflow created with n8n, a no-code automation tool. This process strictly complies with content policies and contains no illegal, offensive, or protected elements. All handled data is legal and publicly accessible.