Turn inbound email into fast, clear replies with a simple review step. Ideal for support teams that need quick answers without losing control. New messages are read, summarized, drafted, and sent only after approval.

Incoming emails are captured over IMAP from Gmail or Outlook. The message is converted to Markdown so the model can read it clearly, then a summarization chain creates a short brief in about 100 words. An agent can reference your business knowledge to keep answers on brand and policy. A writing node produces a professional reply under 100 words, and a reviewer receives both the original email and the draft for a quick check. The final email uses Re plus the original subject and goes back to the sender through SMTP using your mailbox details.

You only need IMAP and SMTP access plus API keys for the language models. Most teams cut reply prep from 10 minutes to 1 minute and keep a steady tone across many agents. Great fit for support queues, shared mailboxes, and any team that wants AI assistance with a human in the loop.

• IMAP listener reads new emails from Gmail or Outlook inboxes
• Markdown node turns HTML emails into clean, readable text for the model
• Summarization chain creates a 100 word brief of the incoming message
• Agent can use your company knowledge so replies match policy and tone
• Reply writer composes a professional answer under 100 words
• Approval email sends both the original message and the draft to a reviewer and waits for a decision
• If check gates the final send based on approval status
• SMTP sender replies with Re plus the original subject and correct to and from fields

• Reduce manual reply prep from 10 minutes to 1 minute
• Automate up to 70 percent of responses with human approval
• Improve consistency and cut wording mistakes by 60 percent
• Handle five times more inbound email without extra staff
• Connect Gmail or Outlook with AI in one streamlined flow
• Keep quality control with a clear approval checkpoint

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 Gmail, Outlook, OpenAI and DeepSeek. See the Tools Required section above for links to create accounts with these services.
3. In the n8n credentials manager, create an IMAP credential for your support mailbox. For Gmail or Outlook, enable IMAP in the account settings and use an app password if two factor authentication is on.
4. Open the Email Trigger node and select your IMAP credential. Choose the Inbox folder and set it to pull unread messages so only new emails are processed.
5. Create an SMTP credential in the n8n credentials manager for the same mailbox. Test the connection, then select it in both the Approve Email and Send Email nodes.
6. Create an OpenAI credential with your API key from the official API page. If you are using a compatible model like DeepSeek, add a separate credential and set the correct base URL if required by your provider.
7. Open the OpenAI and OpenAI Chat Model nodes and choose the matching credentials. Pick the models you want to use for summarization and drafting.
8. In the Approve Email node, set the reviewer email address for your team. Keep send and wait enabled so the workflow pauses for the decision.
9. Check the Write email node system prompt and the 100 word limit. Adjust tone or length rules to match your brand voice.
10. Turn on the workflow. Send a test email to the monitored inbox. Confirm that a summary and draft are created, the approval email arrives, and approving it triggers the final reply to the original sender.
11. If the trigger does not fire, verify IMAP is enabled and the mailbox has unread messages. If approval links do not work, confirm your n8n instance is reachable on the public URL and that webhooks are allowed.