Gmail + Slack: reviewed AI drafts for faster replies
Your inbox fills up, the “quick reply” turns into a 20-minute rewrite, and somehow the important messages get handled last. Manual email replies are slow in a way that’s hard to measure until you’re already behind.
If you run customer support, you feel this daily. A freelance consultant feels it differently, but it’s the same bottleneck. This Gmail Slack automation turns new, unreplied emails into polished AI reply drafts sent to Slack, so you can approve and reply faster without handing control to a bot.
You’ll see what the workflow does, what you need, and how the pieces fit together so you can start getting clean drafts in minutes.
How This Automation Works
Here’s the complete workflow you’ll be setting up:
n8n Workflow Template: Gmail + Slack: reviewed AI drafts for faster replies
flowchart LR
subgraph sg0["Schedule Flow"]
direction LR
n0@{ icon: "mdi:play-circle", form: "rounded", label: "Schedule Trigger", pos: "b", h: 48 }
n1@{ icon: "mdi:cog", form: "rounded", label: "Limit - Max emails per run", pos: "b", h: 48 }
n2@{ icon: "mdi:message-outline", form: "rounded", label: "Gmail - Get message details", pos: "b", h: 48 }
n3@{ icon: "mdi:message-outline", form: "rounded", label: "Search unreplied emails", pos: "b", h: 48 }
n4@{ icon: "mdi:swap-horizontal", form: "rounded", label: "Check if new emails exist", pos: "b", h: 48 }
n5["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/code.svg' width='40' height='40' /></div><br/>Extract email body for AI"]
n6@{ icon: "mdi:robot", form: "rounded", label: "AI - Draft professional emai..", pos: "b", h: 48 }
n7["<div style='background:#f5f5f5;padding:10px;border-radius:8px;display:inline-block;border:1px solid #e0e0e0'><img src='https://flowpast.com/wp-content/uploads/n8n-workflow-icons/slack.svg' width='40' height='40' /></div><br/>Notify draft reply in Slack"]
n8@{ icon: "mdi:message-outline", form: "rounded", label: "Mark email as AUTO_REPLIED", pos: "b", h: 48 }
n0 --> n3
n3 --> n4
n4 --> n1
n5 --> n6
n1 --> n2
n2 --> n5
n2 --> n8
n6 --> n7
end
subgraph sg1["Flow 2"]
direction LR
n9@{ icon: "mdi:swap-horizontal", form: "rounded", label: "OPTIONAL – Auto-send reply", pos: "b", h: 48 }
end
%% Styling
classDef trigger fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
classDef ai fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
classDef aiModel fill:#e8eaf6,stroke:#3f51b5,stroke-width:2px
classDef decision fill:#fff8e1,stroke:#f9a825,stroke-width:2px
classDef database fill:#fce4ec,stroke:#c2185b,stroke-width:2px
classDef api fill:#fff3e0,stroke:#e65100,stroke-width:2px
classDef code fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
classDef disabled stroke-dasharray: 5 5,opacity: 0.5
class n0 trigger
class n6 ai
class n4,n9 decision
class n5 code
classDef customIcon fill:none,stroke:none
class n5,n7 customIcon
Why This Matters: Inbox backlogs slow down revenue and support
Replying to email sounds simple until you’re doing it all day. You read a message, hunt for context, rewrite your first draft to sound “more professional,” then second-guess the tone because you’re representing a brand (or your own reputation). Multiply that by a busy inbox and you get a quiet kind of chaos: replies go out late, customers follow up twice, and your team starts making inconsistent promises because everyone is rushing. It’s not just time. It’s the mental load of switching between “understand the request” and “write the response” over and over.
The friction compounds. Here’s where it breaks down in real life.
- Most replies involve re-reading the email thread and manually pulling details into your response, which wastes a surprising amount of time.
- When multiple people touch the same inbox, you get tone drift and duplicated work because there is no consistent draft process.
- Rushing leads to small errors: wrong names, missed questions, or vague answers that trigger another back-and-forth.
- Auto-send email bots feel risky, so you either avoid automation entirely or spend time cleaning up after it.
What You’ll Build: AI reply drafts in Slack, with human approval
This workflow watches your Gmail inbox for new emails that haven’t been replied to yet. On a schedule, it pulls in a small batch of messages (so you don’t get flooded), fetches the full email content, and extracts the part that matters for a response. Then an OpenAI-powered AI agent generates a polite, professional reply draft based on what the sender asked. Instead of sending anything automatically, the workflow posts the draft to a Slack channel where a human can review it, tweak it, and decide what to send. Finally, the email gets labeled in Gmail so it won’t be processed twice the next time the workflow runs.
The workflow starts with a scheduled inbox check in Gmail. From there, it fetches each message, parses the body, and generates a reply draft using an OpenAI Chat Model. Slack becomes your approval queue, which keeps the speed of automation while preserving your judgement.
What You’re Building
| What Gets Automated | What You’ll Achieve |
|---|---|
|
|
Expected Results
Say you handle about 30 inbound emails a day and each one takes maybe 8 minutes to read, draft, and sanity-check before you send. That’s roughly 4 hours gone. With this workflow, the inbox check happens automatically, the AI creates drafts in the background, and you review in Slack in about 2 minutes per email for the routine ones. You’re still “in the loop,” but you can realistically get about 3 hours back on a busy day.
Before You Start
- n8n instance (try n8n Cloud free)
- Self-hosting option if you prefer (Hostinger works well)
- Gmail for monitoring and labeling inbound email.
- Slack to review drafts and approve replies.
- OpenAI API key (get it from the OpenAI dashboard).
Skill level: Beginner. You’ll connect accounts, paste an API key, and adjust a couple of settings like schedule and Slack channel.
Want someone to build this for you? Talk to an automation expert (free 15-minute consultation).
Step by Step
Scheduled inbox check. A Schedule Trigger runs every few minutes (you pick the timing) and asks Gmail for emails that are new and unreplied.
Filter and cap the workload. If there are no messages, the workflow stops. If there are messages, it limits how many it will process in one cycle so your Slack channel doesn’t get spammed during a surge.
Fetch the full message and extract the useful text. Gmail provides the raw message, then a parsing step pulls out the email body so the AI model gets clean input, not a messy thread full of signatures and formatting.
Generate draft + route to Slack. The OpenAI Chat Model creates a professional reply draft, and the workflow posts it into your chosen Slack channel so someone can review and send the final response.
You can easily modify the schedule frequency to match your inbox volume, and adjust the AI prompt to match your brand tone. See the full implementation guide below for customization options.
Step-by-Step Implementation Guide
Step 1: Configure the Scheduled Trigger
This workflow starts on a timed schedule, so set up the polling interval to check for unread emails.
- Add the Scheduled Email Poller node as your trigger.
- Set the schedule rule to run every
30minutes using the interval configuration. - Connect Scheduled Email Poller to Find Unread Messages.
Step 2: Connect Gmail
Next, configure Gmail access for searching, retrieving, and labeling messages.
- Open Find Unread Messages and set Operation to
getAll. - Set Limit to
10and Filters → Query to=in:inbox is:unread. - Credential Required: Connect your gmailOAuth2 credentials.
- Open Fetch Gmail Message and set Operation to
get. - Set Message ID to
={{ $json.id }}. - Credential Required: Connect your gmailOAuth2 credentials.
- Open Tag Message as AutoReplied and set Operation to
addLabels. - Set Message ID to
={{ $json.id }}and Label IDs to your Gmail label ID (replace[YOUR_ID]). - Credential Required: Connect your gmailOAuth2 credentials.
Step 3: Validate and Limit the Email Batch
Before fetching message details, check that unread emails exist and cap processing volume.
- In Validate Message Presence, set the condition to check if leftValue is
={{$items().length}}and rightValue is0with the operatorgt. - Connect Validate Message Presence to Cap Emails Per Cycle.
- Set Cap Emails Per Cycle Max Items to
5. - Connect Cap Emails Per Cycle to Fetch Gmail Message.
- Ensure Fetch Gmail Message outputs to both Parse Message Content and Tag Message as AutoReplied in parallel.
Step 4: Set Up the Message Parser and AI Draft
Extract the body text from Gmail payloads and generate a polite draft reply in Japanese.
- Open Parse Message Content and keep the JavaScript as provided to decode and clean the email body.
- Confirm the output fields include
emailText,subject,threadId, andreceivedAt. - Connect Parse Message Content to Generate AI Reply Draft.
- In Generate AI Reply Draft, set Model to
gpt-4.1-mini. - In the user prompt, ensure the email body is referenced as
{{ $json.emailText }}. - Credential Required: Connect your openAiApi credentials.
Step 5: Configure Slack Output
Send the AI-generated draft to Slack for review before manual sending.
- Open Post Draft to Slack and set Text to
=📩 There is an email that needs to be replied to\n\n[AI reply draft]\n{{ $node["Generate AI Reply Draft"].json.output[0].content[0].text }}\n\n――――――――――\n(This message is automatically generated.)\n. - Set Select to
channeland choose your Slack channel ID (replace[YOUR_ID]). - Credential Required: Connect your slackApi credentials.
Step 6: Review Optional Auto Send Logic
The workflow includes a placeholder for optional auto-send logic if you later want to send replies automatically.
- Open Utility: Optional Auto Send and review the empty condition fields.
- Leave it disconnected or add your own conditions when you’re ready to automate sending.
Step 7: Test and Activate Your Workflow
Validate the end-to-end flow and then enable it for production use.
- Click Execute Workflow to test manually after ensuring unread emails exist in Gmail.
- Confirm that Find Unread Messages returns items, Generate AI Reply Draft outputs content, and Post Draft to Slack posts the draft.
- Check Gmail to ensure Tag Message as AutoReplied applied the correct label.
- When satisfied, toggle the workflow to Active to run on the schedule.
Troubleshooting Tips
- Gmail credentials can expire or need specific permissions. If things break, check the connected Google account in n8n’s Credentials first, then confirm Gmail API access is still allowed.
- Slack messages sometimes fail when the draft is too long for your formatting. If drafts get cut off, shorten the AI output or post the draft as a file/snippet instead of a single message.
- OpenAI API errors are often quota or model access issues. Double-check your OpenAI billing status and make sure the model selected in the OpenAI Chat Model node is available to your account.
Quick Answers
About 10–15 minutes if Gmail, Slack, and your API key are ready.
No. You’ll connect Gmail and Slack, then paste your OpenAI key and pick a channel.
Yes. n8n has a free self-hosted option and a free trial on n8n Cloud. Cloud plans start at $20/month for higher volume. You’ll also need to factor in OpenAI API costs, which are usually a few cents per day for small inboxes.
Two options: n8n Cloud (managed, easiest setup) or self-hosting on a VPS. For self-hosting, Hostinger VPS is affordable and handles n8n well. Self-hosting gives you unlimited executions but requires basic server management.
Yes, and you probably should. You can adjust the schedule in the Scheduled Email Poller, change which emails qualify in Find Unread Messages, and tweak the wording in Generate AI Reply Draft to match your tone. Many teams also customize the Slack message to include sender, subject, and a direct Gmail link for faster review.
Usually it’s expired Google authorization or the wrong Gmail scope. Reconnect your Gmail credential in n8n, then rerun the workflow and confirm it can list messages and fetch full email content. If it works for a bit and then fails, check for Google security prompts on the account and confirm the mailbox isn’t hitting API limits during high volume runs.
It’s designed to handle steady inbox flow by limiting how many messages are processed per run, so it won’t overwhelm your team.
Often, yes, because you can cap batches, add conditional logic, and fine-tune the parsing without paying extra for every branch. n8n is also easier to keep predictable when you want a human approval loop and you don’t want auto-send mistakes. Zapier or Make can be faster to set up for very simple “Gmail to Slack” forwarding, but the moment you add AI drafts, labeling, and safeguards, the workflow tends to get expensive or brittle. Frankly, the control you get in n8n is the point here. Talk to an automation expert if you’re not sure which fits.
This is the best kind of automation: faster replies without giving up control. Set it up once, then let Slack become the calm, reviewable inbox your team actually sticks to.
Need Help Setting This Up?
Our automation experts can build and customize this workflow for your specific needs. Free 15-minute consultation—no commitment required.
