AI Workflow for Email Reply Drafting

Difficulty: Intermediate · Setup time: 35 min · Steps: 4

Responding to emails is one of the biggest productivity drains in any business — especially for common requests that follow predictable patterns. This workflow monitors your inbox for emails matching defined categories, drafts a professional reply using OpenAI, and prompts you in Slack to review and send with a single click. You go from triage to sent in 30 seconds.

What to Expect

  • 30 sec — from email received to draft ready for review
  • 60% — reduction in time spent on routine email replies

Tools Used in This Workflow

  • Gmail — Incoming email monitor
  • OpenAI — Draft reply generation
  • HubSpot — Email history & CRM context
  • Slack — Review & approve draft

How It Works — Step by Step

  1. Step 1: Monitor Gmail for new emails

    Tool: Gmail

    n8n monitors your Gmail inbox using OAuth and triggers on new emails matching specific criteria — emails from known contacts, emails with certain subject line keywords, or all emails in a specific label. The email subject, body, and sender details are extracted.

  2. Step 2: Fetch CRM context from HubSpot

    Tool: HubSpot

    Using the sender's email address, n8n looks up the contact in HubSpot and fetches their name, company, open deals, and recent activity notes. This context is passed to OpenAI so the reply is appropriately tailored to the relationship.

  3. Step 3: Generate reply draft with OpenAI

    Tool: OpenAI

    OpenAI receives the original email, the CRM context, and your writing guidelines (tone of voice, common phrases to use, any specific instructions for this email category). The output is a complete email reply ready to send, with the subject line pre-filled.

  4. Step 4: Review and approve via Slack

    Tool: Slack

    A Slack message shows you the AI-drafted reply with an Edit and Send link. One click opens the draft in Gmail with all fields pre-filled. You review in seconds, make any edits, and hit Send — no starting from a blank page.

Pro Tips for This Workflow

  • Create a Gmail label called 'AI Draft' and add rules to apply it to emails from key domains — the automation only fires for labeled emails, giving you control over which messages get AI-drafted.
  • Pass the last 3 emails in the thread (not just the latest) to OpenAI — thread context dramatically improves reply coherence and prevents the model from asking questions already answered in the conversation.
  • Add a 'Do Not Draft' label in Gmail and filter it out in n8n — some emails (legal, confidential, complaints) should always go directly to human composition without AI assistance.

What You'll Need

  • Gmail account — Incoming email monitor
  • OpenAI account — Draft reply generation
  • HubSpot account — Email history & CRM context
  • Slack account — Review & approve draft
  • Basic automation experience, or a vetted expert to build it for you

Related Video Tutorial

Auto-Draft Email Replies with AI — Save 1 Hour Per Day

Frequently Asked Questions

Can I configure different reply tones for different email categories?

Yes — add a classification step (either a Zapier filter or a first OpenAI call) that categorizes the email as 'Sales', 'Support', 'Partner', or 'General'. Route each category to a different system prompt with tone instructions (e.g., Sales: warm and enthusiastic; Support: empathetic and solution-focused; Partner: professional and concise).

How do I prevent the AI from making commitments I haven't approved?

In the system prompt, include the instruction: 'Never commit to specific dates, prices, deliverables, or promises — leave placeholders like [DATE TBD] or [PRICE TBD] if these need to be confirmed.' Also add a Slack review step — the draft goes to Slack for approval before it can be sent, so a human always checks before commitments are made.

Does this workflow work with Outlook instead of Gmail?

Yes — replace the Gmail trigger and send steps with Microsoft Outlook nodes in n8n (using the Microsoft Graph API) or use Zapier's Outlook integration. The HubSpot context-fetching and OpenAI drafting steps are identical. Outlook's 'Draft' folder works the same way as Gmail drafts for the review step.

How do I keep the AI draft quality improving over time?

Log sent emails (subject, original incoming email, AI draft, final sent version) to Airtable monthly. Compare AI draft vs. final sent version — where they differ significantly, update your system prompt with the corrected phrasing as an example. After 3–4 months of this iteration, draft quality typically reaches 85%+ approval rate without edits.

Want This Workflow Built for You?

Our vetted automation engineers can set up this exact workflow in days — customized to your tools and process. Free consultation, honest quote, fast delivery.

Get matched free — no commitment

More AI-Powered Workflows workflows | All workflow guides | Find automation experts