Automated FAQ Generator from WhatsApp Groups using GPT-4 and Google Docs

1. Workflow Overview

This workflow automates the generation of Frequently Asked Questions (FAQ) from WhatsApp group conversations using GPT-4 and Google Docs. It is designed to assist educational or community managers in extracting recurring questions and concerns from chat logs, summarizing them with AI, and documenting the results in a Google Doc for easy sharing and reference.

Target Use Cases:

Educational communities seeking to identify common student questions.
Community managers wanting to automate FAQ updates from chat data.
Teams using WhatsApp groups for communication who want structured insights.

Logical Blocks:

1.1 Scheduled Data Retrieval: Triggered weekly to fetch WhatsApp message logs from Google Sheets filtered by specific group IDs.
1.2 Message Grouping and Preprocessing: Using JavaScript code, messages are parsed and grouped by ISO week number, creating message blocks per week.
1.3 AI Analysis: An AI Agent (LangChain) processes each weekly message block, identifying recurring questions and generating a structured FAQ summary.
1.4 Document Preparation: A Google Docs template is copied to create a new document for the current FAQ.
1.5 FAQ Document Update: The AI-generated FAQ content is inserted into the new Google Doc, finalizing the report.

2. Block-by-Block Analysis

2.1 Scheduled Data Retrieval

Overview:
This block triggers the workflow every Monday at 6 AM and retrieves WhatsApp group message data from a Google Sheets document filtered by group identifiers.
Nodes Involved:
- Schedule Trigger
- Get row(s) in sheet
Node Details:
- Schedule Trigger
  - Type: n8n Schedule Trigger
  - Configuration: Runs weekly on Monday at 6:00 AM.
  - Inputs: None (start node).
  - Outputs: Triggers the next node.
  - Edge Cases: Timezone misconfigurations could cause unexpected trigger times. Ensure server timezone matches expectations.
- Get row(s) in sheet
  - Type: n8n Google Sheets node
  - Configuration: Reads rows from a Google Sheet identified by document ID "1WNX4TNB0AEFXGnBbGYJqz0by9TcHcUd5Z5cH4GsGZ6g" (sheet "gid=0").
  - Filters: Retrieves rows where "RemoteJid Grupo" column matches either "120363418264921236@g.us" OR "120363401761323889@g.us" (two WhatsApp group IDs).
  - Credentials: Uses Google Sheets OAuth2 credential "Google Sheets - Luís".
  - Inputs: Trigger from Schedule Trigger.
  - Outputs: List of message rows matching the filter.
  - Edge Cases:
    - API rate limits or OAuth token expiration may cause failures.
    - Empty or malformed data rows may propagate errors downstream.

2.2 Message Grouping and Preprocessing

Overview:
Transforms raw messages into weekly grouped blocks by parsing message dates and aggregating messages per week. Prepares data for AI processing.
Nodes Involved:
- Code
Node Details:
- Code
  - Type: n8n Code node (JavaScript)
  - Configuration:
    - Parses dates from "Data" column (format dd/MM/yyyy).
    - Computes ISO week number for each message date.
    - Groups messages by week, concatenates all messages into a single text block per week.
    - Outputs an array of objects with properties: "semana" (week string) and "bloco" (concatenated messages).
  - Input: Rows from Google Sheets.
  - Output: Weekly grouped message blocks (one per week).
  - Edge Cases:
    - Missing or malformed dates ("Data") are skipped.
    - Incorrect date formats lead to skipped messages.
    - Very large message blocks may exceed downstream API limits.
  - Notes: The code uses ISO week calculation to standardize weeks.

2.3 AI Analysis

Overview:
Uses a LangChain AI Agent configured with a system prompt to identify recurring doubts and questions from message blocks. Produces a structured FAQ in text format.
Nodes Involved:
- AI Agent
- OpenAI Chat Model
Node Details:
- AI Agent
  - Type: LangChain AI Agent node
  - Configuration:
    - Receives input: text containing the week ID and concatenated message block.
    - System Message describes the task: identify recurring questions, explanations, and suggested answers for FAQs.
    - Output: organized, clear list of recurring questions with explanations and optional suggested responses.
  - Input: Weekly message blocks from Code node.
  - Output: FAQ text summary per week.
  - Edge Cases:
    - Overly long message blocks might cause token limit errors.
    - Misinterpretation of ambiguous messages may lead to incomplete or incorrect FAQs.
    - AI service availability and authentication issues are possible.
  - Sub-Workflow: Utilizes the OpenAI Chat Model node as the language model backend.
- OpenAI Chat Model
  - Type: LangChain OpenAI Chat Model node
  - Configuration: Uses GPT-4.1-mini model (GPT-4 variant).
  - Credentials: OpenAI API key "OpenAi account".
  - Input: Receives prompt from AI Agent.
  - Output: AI-generated text passed back to AI Agent.
  - Edge Cases:
    - API rate limits, quota, or network errors may interrupt processing.
    - Model version must support GPT-4.

2.4 Document Preparation

Overview:
Copies a Google Docs template document to create a new document where the FAQ will be written.
Nodes Involved:
- Copy file
Node Details:
- Copy file
  - Type: n8n Google Drive node
  - Operation: Copy a file.
  - Configuration: Copies the template document with ID "1hAStPci4ZgaXMPbpLRh4U8MxeJUIkul_iYwYzhQVaus".
  - New file named: "Extraction FAQ - Whatsapp Group".
  - Credentials: Google Drive OAuth2 "Google Drive account".
  - Input: Triggered after AI Agent produces output.
  - Output: Metadata of the new copied Google Doc, including its URL and ID.
  - Edge Cases:
    - Google Drive API errors or permission issues.
    - Template document changes would affect output format.

2.5 FAQ Document Update

Overview:
Inserts the AI-generated FAQ content into the newly created Google Doc.
Nodes Involved:
- Update a document
Node Details:
- Update a document
  - Type: n8n Google Docs node
  - Operation: Update document by inserting text.
  - Configuration: Inserts the text output from the AI Agent node into the new Google Doc identified by its document URL (from Copy file output).
  - Credentials: Google Docs OAuth2 "Google Docs account".
  - Input: Receives new document ID and AI-generated FAQ text.
  - Edge Cases:
    - Document permission or API quota issues.
    - Large text insertion might cause timeout or truncation.

3. Summary Table

Node Name	Node Type	Functional Role	Input Node(s)	Output Node(s)	Sticky Note
Schedule Trigger	n8n-nodes-base.scheduleTrigger	Triggers workflow weekly	None	Get row(s) in sheet	Run Every Week: Every Monday 6am
Get row(s) in sheet	n8n-nodes-base.googleSheets	Retrieves WhatsApp messages from Google Sheets	Schedule Trigger	Code	Find all messages of that last week
Code	n8n-nodes-base.code	Groups messages by ISO week	Get row(s) in sheet	AI Agent	Do "messages blocks" and identify the week of those messages
AI Agent	@n8n/n8n-nodes-langchain.agent	Analyzes messages block and generates FAQ text	Code	Copy file	The AI Agent separates the questions and the context about these questions
OpenAI Chat Model	@n8n/n8n-nodes-langchain.lmChatOpenAi	Provides GPT-4 language model for AI Agent	AI Agent (ai_languageModel)	AI Agent
Copy file	n8n-nodes-base.googleDrive	Copies FAQ template document	AI Agent	Update a document	Copy the template doc. file and create new after that
Update a document	n8n-nodes-base.googleDocs	Inserts AI-generated FAQ text into Google Doc	Copy file	None

4. Reproducing the Workflow from Scratch

Create Schedule Trigger node:
- Type: Schedule Trigger
- Set to run weekly: every Monday at 6:00 AM.
Create Google Sheets node ("Get row(s) in sheet"):
- Type: Google Sheets node (read operation)
- Connect input from Schedule Trigger.
- Set Document ID to your WhatsApp messages spreadsheet ID.
- Set Sheet Name to the correct sheet (gid=0).
- Add filter to "RemoteJid Grupo" column: include rows where value equals either "120363418264921236@g.us" OR "120363401761323889@g.us".
- Use your Google Sheets OAuth2 credentials.
Create Code node:
- Connect input from Google Sheets node.
- Paste JavaScript code that:
  - Parses message dates in dd/MM/yyyy format.
  - Converts dates to ISO weeks.
  - Groups messages by week and concatenates them into a single string per week.
- Output: JSON objects with "semana" (week) and "bloco" (message block).
Create LangChain AI Agent node:
- Connect input from Code node.
- Configure prompt inputs to include:
  - "Semana que está sendo analisada: {{ $json.semana }}"
  - "Bloco de Mensagens a ser tratado e analisado: {{ $json.bloco }}"
- Set System Message to instruct the AI to identify recurring questions, explain them, and provide suggested answers, as per the detailed prompt in the original workflow.
- Set to use "define" prompt type.
Create OpenAI Chat Model node:
- Use GPT-4.1-mini model.
- Use your OpenAI API credentials.
- Connect as AI language model backend of the AI Agent node.
Create Google Drive node ("Copy file"):
- Connect input from AI Agent output.
- Operation: Copy a file.
- Set File ID to your FAQ template Google Doc ID.
- Name the new file appropriately (e.g., "Extraction FAQ - Whatsapp Group").
- Use Google Drive OAuth2 credentials.
Create Google Docs node ("Update a document"):
- Connect input from Copy file node.
- Operation: Update document by inserting text.
- Insert text from AI Agent's output JSON (FAQ content).
- Document URL set dynamically from Copy file node output.
- Use Google Docs OAuth2 credentials.
Connect nodes in order:
Schedule Trigger → Get row(s) in sheet → Code → AI Agent → Copy file → Update a document.
Test the flow:
- Run manually or wait for scheduled trigger.
- Verify Google Docs file creation and content correctness.

5. General Notes & Resources

Note Content	Context or Link
Workflow runs every Monday at 6 AM to produce weekly FAQ updates from WhatsApp group chats.	Time scheduling information from sticky notes.
Source WhatsApp messages are stored in a Google Sheet filtered by group IDs (WhatsApp group JIDs).	Group IDs filter: "120363418264921236@g.us", "120363401761323889@g.us".
AI Agent prompt is carefully designed to extract recurring questions with explanations and suggestions.	Detailed system prompt included within AI Agent node parameters.
Google Docs template document is copied to ensure consistent formatting and to keep a master template intact.	Template Doc ID: "1hAStPci4ZgaXMPbpLRh4U8MxeJUIkul_iYwYzhQVaus".
OpenAI GPT-4 model used via LangChain integration requires valid OpenAI API credentials.	Ensure API key with GPT-4 access and sufficient quota.
Google OAuth2 credentials for Sheets, Drive, and Docs must have proper scopes and permissions set.	Credentials setup required for Google APIs: Sheets (read), Drive (copy), Docs (update).
Edge cases include handling missing dates, API rate limits, AI prompt length limits, and document write failures.	Plan for retries, error handling, and message size limitations.
The workflow is inactive by default (active=false) and should be activated after credential verification and testing.	Activation step required in n8n UI before scheduling begins.

Disclaimer:
The text provided is extracted exclusively from an automated workflow created with n8n, a tool for integration and automation. This processing strictly complies with applicable content policies and contains no illegal, offensive, or protected elements. All data handled is legal and public.

14 KiB Raw Blame History