Handle incoming info requests without manual typing. The flow reads emails, checks if they ask about your company, and sends a short reply based on your internal files. It fits shared inboxes like info and pre sales questions.

A setup path builds a private knowledge base from your Google Drive. It pulls documents from a chosen folder, splits the text, creates embeddings with OpenAI, and stores them in Qdrant. For each email, the content is converted from HTML to Markdown, summarized with a smart model, and classified. If the message is a general company info request, an agent drafts a reply using Qdrant search as a tool, a second model reviews the text into clean HTML, and the message is sent back with the original subject line. Emails that do not match are skipped.

You need access to your mail server with IMAP and SMTP, a Google Drive folder for your documents, a Qdrant endpoint, and API keys for OpenAI and OpenRouter. Run the setup to index files, then process your inbox. Teams can cut response time, improve answer quality, and handle more volume without adding headcount.

• Reads emails via IMAP and passes content to the flow
• Converts HTML email to Markdown for better model reading
• Summarizes long emails to a short brief for faster decisions
• Classifies messages and only continues for company info requests
• Uses Qdrant retrieval so the model answers from your documents
• Builds a knowledge base from a Google Drive folder with token splitting
• Creates document embeddings using OpenAI for accurate search
• Creates and refreshes Qdrant collections through HTTP API calls
• Reviews and formats the reply into clean HTML before sending
• Sends the response by SMTP with correct to and subject mapping

• Reduce manual inbox replies from 2 hours daily to 10 minutes
• Automate about 70 percent of general info emails
• Improve reply consistency by 90 percent with a review step
• Connect email, Google Drive, and Qdrant in one flow
• Handle 5 times more messages without adding staff
• Cut first response time to under 1 minute after indexing

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Google Drive, Qdrant, OpenAI, OpenRouter, IMAP Email and SMTP Email. See the Tools Required section above for links to create accounts with these services.
3. In the n8n credentials manager, create IMAP Email credentials. Enter your mail host, username, password, port, and SSL settings. Test the connection.
4. Create SMTP Email credentials. Add host, port, username, password, and secure settings. Send a test email from n8n to confirm.
5. Connect Google Drive with OAuth. In the Get folder node, select your credential and choose the target folder that holds company docs.
6. Set up Qdrant. Create an HTTP Header Auth credential with your API key. In the Create collection and Refresh collection nodes, replace QDRANTURL and COLLECTION with your values. Also set the Qdrant credential in both Vector Store nodes and confirm the same collection name.
7. Add your OpenAI API key. Attach it to the Embeddings nodes and any OpenAI LLM nodes. Keep the chosen models as provided or select similar ones.
8. Add your OpenRouter API key. Attach it to the DeepSeek LLM nodes used for summarization and review.
9. Open the Email Classifier node and confirm the category list includes Company info request. Adjust the prompt or categories if your use case needs more options.
10. Run the manual setup path. Click Test workflow to create and populate the Qdrant collection from Google Drive. Check execution logs for successful document loading and embeddings.
11. Send a test email to the connected inbox. Run the workflow and confirm the message is summarized, classified, answered from the knowledge base, reviewed into HTML, and sent back with Re plus the original subject.
12. If emails are not read, verify IMAP access and mailbox permissions. If retrieval fails, ensure the Qdrant URL, API key, and collection name match across all nodes.