Automate Company Document Queries with Google Drive & Pinecone

What This Automation Does

This workflow watches a Google Drive folder for new or changed files.

It downloads updated documents, breaks them into small parts, makes embeddings with Google Gemini, and saves them in Pinecone.

When someone asks a question in chat, it finds answers in those documents fast.

This stops slow manual searching and helps answer questions quickly and right.

Workflow Inputs, Processing, and Outputs

Inputs

• Google Drive Folder: Documents stored and updated here trigger the workflow.


• User Questions: Natural language queries sent through a chat interface.

Processing Steps

• Trigger when a file is created or updated in the Google Drive folder.


• Download the document using its file ID.


• Split document text into overlapping chunks for better search context.


• Generate semantic embeddings of text chunks with Google Gemini’s text-embedding model.


• Insert these vectors into Pinecone index named company-files for fast lookup.


• On chat message, query the Pinecone index to find relevant document chunks.


• Use Google Gemini’s chat model to generate a helpful answer based on those chunks.


• Keep recent chat messages in a window buffer memory to remember conversation flow.

Outputs

• Instant, accurate answers to employee questions from company documents.


• Automated document indexing without manual work.


• Improved HR productivity and faster onboarding support.

Who Should Use This Workflow

This workflow is good for HR teams or company staff who want quick answers to questions about policies, benefits, or procedures.

If your documents are on Google Drive and employee questions take too long to answer, this workflow helps.

It fits any company size that uses Google Drive and wants to keep a live searchable knowledge base.

Tools and Services Used

• Google Drive: Stores company documents.


• n8n automation platform: Runs the workflow automation.


• Google Gemini (PaLM) API: Creates text embeddings and chat answers.


• Pinecone vector store: Stores and searches vector embeddings in index “company-files”.

Beginner Step-by-Step: How to Use This Workflow in n8n

Step 1: Import the Workflow

1. Download the workflow JSON file using the Download button on this page.


2. In n8n editor (assumed already open), go to “Import from File”.


3. Select the downloaded JSON file to import the workflow.

Step 2: Add Required Credentials

1. Set up Google Drive OAuth2 credentials with the right account and folder access.


2. Add Google Gemini API credentials with API Key.


3. Add Pinecone API Key and configure connection to the “company-files” index.

Step 3: Configure IDs and Parameters

1. Update the Google Drive Trigger nodes with the correct folder ID to watch.


2. Check the Pinecone node is pointing to the “company-files” index.


3. If needed, adjust chunk overlap parameter in the Recursive Text Splitter node.

Step 4: Test the Workflow

1. Manually upload or update a test file in the Google Drive folder.


2. Run the workflow manually or watch the trigger pick it up.


3. Send a test chat question through the configured webhook or chat interface.


4. Confirm a relevant answer is returned from the AI Agent.

Step 5: Activate for Production

1. Turn on the workflow to start automatic processing.


2. Monitor executions and fix errors as needed.


3. For reliable long-term running, consider self-host n8n.

Customization Ideas

• Change the Google Drive folder in trigger nodes to watch different document sets.


• Replace Pinecone with other vector databases like Weaviate or FAISS if preferred.


• Adjust text chunk overlap to balance search relevance and processing speed.


• Edit AI Agent system messages to change tone or add instructions for answers.


• Add multi-language detection to support users speaking different languages.

Common Issues and Fixes

Issue: No documents found or empty search answers

Check Pinecone API keys and index name “company-files”.

Make sure text chunks are generated correctly by the splitter.

Issue: AI Agent says "I cannot find the answer"

Verify vector store tool connects properly to Pinecone retrieval node.

Check embedding and chat nodes have correct API access.

Issue: Workflow not triggering or file download errors

Confirm Google Drive folder ID is set correctly in triggers.

Reauthorize Google Drive OAuth2 credentials in n8n if needed.

Check polling frequency is frequent (e.g., one minute) to avoid delays.

Pre-Production Checklist

• Verify Google Drive folder is correct and accessible with current credentials.


• Test uploading and updating files in the folder to start the workflow.


• Check Pinecone index “company-files” exists with valid API keys.


• Make test chat queries to confirm good AI answer responses.


• Save workflow and credential backups before going live.

Deployment Guide

Activate Google Drive trigger nodes inside n8n.

Make sure the n8n instance runs continuously to catch events.

Use the n8n UI to check workflow run history and fix errors promptly.

Set alerts for failures if possible to stay informed.

For stable use inside a company, consider self-host n8n on a VPS or cloud server.

Summary of Results

→ The Google Drive folder stays automatically indexed in Pinecone.

→ Employees get instant answers to questions from up-to-date documents.

✓ Saves HR teams many hours of searching and manual work.

✓ Keeps the company knowledge base always fresh and accessible.

✓ Improves employee satisfaction with quick and accurate help.