Multi-Modal Appointment Booking Agent for WhatsApp using Cal.com - n8n Workflow

• Businesses or professionals needing 24/7 automated booking support.


• Developers looking for advanced n8n templates integrating LangChain Agents and multi-modal AI.


• Users aiming to connect WhatsApp Business API with Cal.com for scheduling.


• Anyone requiring a complex, stateful n8n automation solution.

Managing appointments across multiple communication channels can be labor-intensive. This powerful n8n workflow solves this by creating a highly intelligent, multi-modal conversational agent directly within WhatsApp. When a customer sends a message (be it text, a voice note, or an image of a calendar), the n8n system processes the input using advanced OpenAI models (Whisper for audio, GPT-4o Vision for images) to extract the intent.

This core capability feeds into an AI Agent n8n node, which uses Cal.com tools to perform real-time scheduling tasks (check availability, book, cancel, reschedule). The entire process is managed within this single n8n workflow, ensuring fast, accurate, and context-aware responses, greatly enhancing customer service efficiency. This is one of the most comprehensive n8n templates available for conversational AI.

The entire process begins with the WhatsApp Trigger, a dedicated n8n trigger that listens for new messages from the Meta Graph API.

1. Message Reception and Routing: The n8n trigger captures the incoming message. The workflow immediately splits the message and uses a Switch n8n node (Identify and ReRoute Message Types) to determine if the message is 'Text', 'Audio', or 'Image'.


2. Multi-Modal Input Processing: Depending on the type, the message is routed:

1. Prompt Preparation: All paths converge into Edit Fields n8n nodes that unify the output into a single userprompt variable, ready for the AI.


2. AI Orchestration (The Core n8n Workflow): The AI Agent n8n node takes the prompt, leveraging the Simple Memory n8n node to recall past conversation history (critical for sequential booking steps). The Agent uses a GPT-4 model and a rigorous System Message, which defines complex French-language policies and tool use rules (e.g., time conversion between Paris and UTC, requiring confirmation before booking, and filtering past appointments). It dynamically selects one of the five Cal.com tool n8n nodes (checkavailability, book_appointment, etc.) to fulfill the user's request.


3. Response Delivery: The final response text generated by the AI Agent passes through a Code n8n node to convert markdown bold text (example) into Unicode bold for optimal WhatsApp display. The Send message n8n node then delivers the professional, automated reply back to the client's WhatsApp number.

To deploy this advanced n8n workflow, follow these steps:

1. Import the n8n template: Copy the provided JSON and import it directly into your n8n instance.


2. Configure WhatsApp Credentials: Set up the required WhatsApp Trigger API and WhatsApp API credentials within n8n. This requires setting up a Meta App and obtaining your Phone Number ID and Access Token.


3. Configure OpenAI Credentials: Set up the OpenAi account credential, as this n8n workflow relies on Whisper (for transcription), GPT-4o (for vision analysis), and GPT-4 (for the AI Agent language model).


4. Configure Cal.com Tools: Update the bookappointment, findbooking, cancelbooking, and reschedulebooking n8n nodes. Replace YOURTOKENHERE in the Authorization headers with your Cal.com API key. Also, verify and update any hardcoded eventTypeId values within the Cal.com tool n8n nodes.


5. System Message Customization: Review the AI Agent n8n node’s System Message to customize the assistant's persona, operational hours, language rules (default is French), and the specific eventTypeId for your service.


6. Activate Webhook: Copy the Webhook URL from the WhatsApp Trigger n8n node and paste it into your Meta Developer Console configuration for the WhatsApp product.


7. Test and Activate: Ensure the n8n trigger is active and test the flow by sending various message types (text, audio, image) to your linked WhatsApp number.

WhatsApp Trigger (n8n trigger): The entry point of the n8n workflow. It monitors the WhatsApp Business API for new incoming messages.
Identify and ReRoute Message Types1 (Switch n8n node): Essential flow control, routing the item based on type (audio, image, or text) to the correct processing pipeline.
Audio Transcriber (OpenAI n8n node): Uses the Whisper model to convert downloaded audio files into a readable text transcript.
Analyze image (OpenAI n8n node): Uses GPT-4o Vision capabilities to describe the image content and extract text, crucial for multi-modal processing.
AI Agent (LangChain Agent n8n node): The core intelligence. It uses the input prompt, conversation history (Simple Memory), and the GPT-4 Chat Model to orchestrate high-level scheduling logic via the Cal.com tools. The detailed System Message dictates its behavior (e.g., French communication, time zone handling).
checkavailability (HTTP Request Tool n8n node): One of the specialized tool n8n nodes. It queries the Cal.com API to find open time slots based on user-requested dates and timezones.
bookappointment / cancelbooking / reschedulebooking (HTTP Request Tool n8n nodes): These are the transactional tool n8n nodes that execute the actual appointment management actions in Cal.com.
Code (n8n node): A post-processing step that ensures the final response is properly formatted for WhatsApp by converting markdown bold (text) to Unicode bold characters, enhancing readability.

• Send message (WhatsApp n8n node): Sends the final, formatted AI response back to the original WhatsApp sender.