How to Summarize Incoming Emails with AI in Make.com

Build a Make scenario that watches a Gmail inbox and posts a short AI summary of each new email to Slack.

8 minBeginner

Long emails eat time. In this guide you will wire Gmail to OpenAI to Slack inside Make.com so that every new message arrives in your channel as a two sentence summary. No code, just three connected modules and one prompt.

A free or paid Make.com account
A Gmail account you can connect (or any IMAP mailbox)
An OpenAI API key from platform.openai.com
A Slack workspace where you can install apps

Step 1: Create the scenario and add the Gmail trigger

From the Make dashboard click Create a new scenario. Search for Gmail and pick the Watch emails module. Connect your Google account, set the folder to Inbox, and set Maximum number of results to 1 so the scenario handles one email per run while you test.

Make.com - scenario editor

[ + ] New scenario

Module 1 Gmail > Watch emails

Connection: my-gmail

Folder: INBOX

Criteria: All emails

Max results: 1

The Gmail Watch emails module is the first node in the flow.

Step 2: Add the OpenAI module

Click the small plus next to the Gmail module and add OpenAI (ChatGPT, Whisper, DALL-E). Choose the Create a Completion (Chat) action. Create a connection by pasting your API key. Pick a model such as gpt-4o-mini, which is cheap and fast enough for summaries.

Step 3: Write the summarize prompt

In the Messages field add one message with role System and one with role User. The system message sets the behavior, and the user message feeds in the email body using the Gmail mapping token. Keep the instruction tight so the model returns clean text with no preamble.

OpenAI message (role: user)

Summarize this email in two short sentences. State who it is from and what action, if any, is requested. Do not add a greeting.

From: {{1.from}}
Subject: {{1.subject}}
Body:
{{1.text}}

Lower your token cost

Set Max Tokens to about 120 and Temperature to 0.2. Summaries do not need creativity, and capping tokens keeps each run a fraction of a cent.

Step 4: Post the summary to Slack

Add a Slack Create a Message module. Pick the target channel, then map the OpenAI output (usually called result or Choices[].Message.Content) into the Text field. Add the original subject above it so the channel message has context.

Slack - #inbox-digest

Agent

Subject: Q3 invoice follow-up Summary: Accounting at Vellum Co asks you to confirm the revised invoice total by Friday. They need a yes or no reply, not a new document.

How a finished summary lands in the channel.

Step 5: Test and turn it on

Click Run once. Make pulls the latest email, runs the prompt, and posts to Slack. Check each module bubble for a green check. When it works, save the scenario and toggle Scheduling on so it runs every 15 minutes.

Result: every new email now appears in Slack as a clean two sentence brief, so you can triage your inbox without opening Gmail.