Skip to content
ChatGPTAppsRank
Submit App

How to Use Slack in ChatGPT

Catch up on channels, search past conversations, and send messages without leaving ChatGPT.

Editorial80Category: BusinessRead the full Slack review →Last tested Jun 10, 2026

Who this guide is for

Anyone whose team runs on Slack and wants ChatGPT to summarize channels, retrieve specific decisions from history, and write to the right place without context-switching.

Why use Slack inside ChatGPT

Slack inside ChatGPT replaces the morning ritual of scrolling channels to figure out what happened overnight. Ask for a summary of #engineering since yesterday, pull every message tagged with a specific keyword, or send a status update to a channel — all from the chat surface. For managers and async-heavy teams, this is one of the highest-leverage connections.

Before you start

  • A Slack workspace where the channels you care about already exist.
  • Workspace owner consent — most Slack plans require an owner to approve the connection.
  • An understanding of which channels are public vs private. Private-channel access requires explicit authorization.

Step-by-step: using Slack inside ChatGPT

  1. Step 1

    Connect Slack

    From the app picker, select Slack. You'll be sent through Slack's OAuth flow where you choose the workspace and scope. Start with read scopes (channels:history, search) and add write later if needed. The connection persists per-workspace until revoked.

  2. Step 2

    Summarize a channel since a timestamp

    The single most-used flow. Ask for a structured catch-up — by topic, by decision, by owner.

    Try this prompt

    Summarize #engineering since 5pm yesterday — group by topic, list any decisions made, and flag anything still open.

    What to expect

    A topic-grouped summary with bullets for each thread, decisions highlighted, and open items called out. Links back to the original Slack messages when supported.

  3. Step 3

    Search past conversations

    When you know a decision was made but can't remember the channel — let the model search across everything you have access to.

    Try this prompt

    Find the conversation where we decided to delay the pricing change. It was in the last two weeks, probably in #leadership or #pricing.

    What to expect

    A snippet of the relevant thread with a link to the original message, plus surrounding context.

  4. Step 4

    Send a message to a channel

    Write scopes enable sending. Always confirm before fire-and-forget — channel-wide messages have real social cost.

    Try this prompt

    Send a short update to #releases: 'Shipping v2.4 today at 2pm PT. Changelog in the release notes doc.' Confirm before sending.

    What to expect

    A preview of the message and a confirmation prompt. Sent only after explicit approval.

  5. Step 5

    Disconnect when the project ends

    If you connected for a specific project, revoke when done. Workspace owner can also revoke at any time from Slack's admin panel.

Common pitfalls

  • Granting full workspace history without thinking. For sensitive workspaces (HR, legal), scope to specific channels only.
  • Letting the model send messages without confirmation. Channel posts are visible to everyone — confirm before send.
  • Trusting summaries blindly. Always click through to source messages on consequential decisions.

Related on ChatGPTAppsRank

Full Slack review (Score 80/100)Best ChatGPT apps for businessSee all rankingsApps vs GPTs explained

Frequently asked questions

Does it work with private channels?
Only if you grant explicit access. The OAuth flow shows which channels the integration can read; private channels require you to add the bot.
Can it use Slack threads?
Yes — threaded replies are part of the channel history and surface in summaries when relevant.
Is Slack content used for training?
Review OpenAI's connector data policy and Slack's data export terms. Enterprise plans typically have explicit no-training commitments; consumer accounts depend on your account settings.