Wiring a Slack bot to your CRM in under an hour

The trigger for this one: a deal closes in the CRM, and the first anyone on the team hears about it is when the invoice bounces because nobody set up the account. A Slack ping the moment the stage changes fixes that, and it takes less setup than most people expect.

This is the same shape of workflow whether you’re watching for a new deal, a status change, or a support ticket sitting untouched too long — one trigger, one lookup, one message. Here’s the version we actually run.

What you’re connecting

• An n8n instance
• Your CRM’s outgoing webhook — or a polling trigger if it doesn’t have one
• A Slack workspace you can install a bot into, with a channel to post to

Trigger: take the webhook over polling, when you can get it

Most CRMs worth using can fire a webhook on record changes — deal stage updates, new contacts, whatever you care about. Use that if it exists. Polling on a schedule works too, but it means checking every few minutes whether anything changed, which is slower to notice and does more work for no benefit.

If your CRM only offers polling, filter aggressively in the node right after the trigger. Don’t let “check every record” turn into “process every record” further down the workflow.

Formatting the message so people actually read it

A wall of raw CRM fields dumped into Slack gets skimmed once and ignored after that. We kept ours to three lines: who, what changed, and a link straight into the record. Slack’s block formatting supports a button on the message itself — we used it to jump directly to the CRM record instead of making anyone go find it.

The gotcha that cost us twenty minutes the first time: Slack message formatting uses its own markup, not standard Markdown — bold is single asterisks, not double. Get that wrong and the message posts with literal asterisks in it instead of bold text, which looks broken enough that people stop trusting the bot.