Get faster, consistent replies to incoming emails without writing them by hand. New messages in your Outlook inbox are picked up, drafted in your voice, and either sent back to the sender or saved as a draft for review. It is ideal for customer support teams, busy founders, or anyone who handles repeat questions.

New emails are detected by an Outlook trigger that checks your mailbox every minute and passes the raw message to an AI agent. The agent uses an OpenAI chat model and your tone samples to write a clear, professional reply. It uses Outlook to place the response on the original thread, reply to the sender only, and keep the subject line stable. You can let the agent send the reply or save it to Drafts for a human checkpoint.

To set it up, connect your Microsoft Outlook and OpenAI accounts, add a few real reply examples to teach tone, and confirm the sender filter if you only want to handle certain mail. Most teams see faster first responses and fewer manual touches within the first day. Common uses include support ticket answers, pre sales questions, and quick follow ups that stay on brand.

• Outlook message received trigger with a sender filter to target only the emails you want handled.
• Polling every minute so new messages are captured quickly without manual checks.
• Raw email payload passed to the agent to preserve key fields like id, from, and subject.
• OpenAI chat model selection such as gpt 4o mini for fast and cost effective replies.
• System prompt with clear rules and your sample replies to match your writing style.
• Outlook Reply tool uses the original messageId to keep the thread intact and replies only to the sender.
• Subject passthrough and replyTo mapping to keep context clean and accurate.
• Optional save to Drafts mode so a human can approve or edit before sending.

• Reduce manual email drafting from 2 hours a day to about 15 minutes
• Automate up to 70% of routine replies with an AI agent trained on your tone
• Improve first response time by up to 60% for inbound messages
• Keep brand voice consistent by using your own reply examples
• Lower risk with an optional Drafts step for human approval
• Connect Outlook and OpenAI in one flow without switching tools

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 Microsoft Outlook and OpenAI. See the Tools Required section above for links to create accounts with these services.
3. Open the Outlook trigger node in n8n and connect your mailbox. In the Credential dropdown, click Create new credential and follow the on screen steps to allow mail read and send permissions.
4. In the Outlook trigger, set the event to message received, choose output raw, and keep the poll time at every minute. Use the sender filter to limit which emails get handled or remove it to cover all mail.
5. Open the OpenAI Chat Model node. In the Credential dropdown, click Create new credential, then paste your OpenAI API key from the OpenAI dashboard and save.
6. Select the model you want, such as gpt 4o mini, and save the node.
7. Open the AI Agent Instructions node. Paste several real reply samples so the agent mirrors your tone. Keep the rules short and clear.
8. Open the Outlook Reply tool node. Confirm operation Reply, map messageId to {{$json.id}}, toggle Reply to sender only, and keep the subject mapped from the incoming email.
9. If you want a human checkpoint, configure an alternate Outlook tool action to save replies to the Drafts folder. Use the same field mappings as the send action.
10. Send a test email that matches your filter. Check Executions in n8n to confirm the trigger fired and see the AI message. Verify a reply was sent or a draft was created in Outlook.
11. If nothing happens, recheck the sender filter, make sure credentials are connected, confirm the model name is valid, and verify the messageId field is mapped to the incoming email id.