Skip to main content
Integration · Slack

Get notified in Slack when your bot needs you

Your bot handles the unglamorous work all day. But some moments still need a human, and they need a human now. A new lead. An escalation. A question your knowledge base doesn’t cover. NebulaHex turns those moments into a single Slack message in the channel your team already lives in.

OAuth in two minutes · Works on Slack Free tier · Per-bot channel routing

Acme Workspace
Channels
# general
# support
# leads1
# founders
# bot-alerts
# leads3 members
Sarah9:42 AM
Following up on yesterday’s outbound
NebulaHexAPPnow
🎯 New lead captured
Priya Sharma· priya@stripe.com
Asked about: enterprise pricing for 50+ seat team
Open conversation
Message #leads
Jump to section

01

Slack is for your team, not your customers

Most chatbot integrations are about giving your bot more places to talk to customers. WhatsApp, Instagram, Messenger, Telegram, the website widget — those are channels where end users send messages and your bot replies. The Slack integration is something else entirely.

Slack is where your teamgets notified about what the bot is doing. It is not a place where your customers will ever interact with your bot — because your customers are not in your Slack workspace. This is a feature, not a limitation. There is nothing useful for a customer-facing bot to do inside Slack.

The conversations that matter to your business — pre-sales questions, support tickets, lead capture — happen on the channels where your customers actually live. Slack sits one layer above all of that, on the team coordination side. It’s where your founders, support team, sales reps, and on-call engineers are already paying attention.

So when we say “Slack integration,” we mean exactly this: when something happens that your team needs to know about, NebulaHex posts a message into a Slack channel of your choosing. The bot does the work; Slack tells you when the work needs you.

02

What NebulaHex sends to Slack

Five event types you can subscribe a Slack channel to. Three are turned on by default because they’re the events most teams care about; two are off by default because they’re high-volume and usually only enabled for specific debugging or audit reasons.

🎯

Lead captured

On

Your bot asked a visitor for their name and email — usually as part of a "talk to sales" or "book a demo" flow — and they handed over details. For founders and small sales teams, this turns a passive widget into a live pipeline tool.

🚨

Conversation escalated

On

A customer asked to speak to a human, or the bot decided the question was outside its lane. The Slack alert is the nudge that gets a human on a waiting conversation faster than a polling refresh of a queue ever would.

💡

Fallback triggered

On

The bot answered, but confidence was low. Each fallback alert is a free signal — a customer asked something your knowledge base doesn’t fully know. Over a week, those alerts become a backlog of pages to write and FAQs to add.

💬

New message

Off

Every inbound customer message gets posted to Slack. For teams that want a real-time firehose — usually founders in the first month after launch, or QA-focused teams spotting bot misbehavior in the wild.

Test completed

Off

When you run an automated test against your bot — checking that an answer still matches expectations after a knowledge update — the result gets posted. Useful for teams where bot quality is owned by a specific person.

You pick which events go to Slack. You pick which channel they land in. You can have only the high-signal three flowing into your main team channel and use the other two only when actively investigating. Or you can route everything to a quieter #bot-firehoseand let a few people watch the whole stream. There’s no wrong configuration; the events are tools for shaping signal-to-noise.

03

Connect in two minutes

Connecting Slack is a single OAuth flow. From your bot’s Integrations tab in the dashboard:

1

Click Connect Slack

From your bot’s Integrations tab in the dashboard, click the Connect Slack button on the Slack card.

2

Authorize on Slack

Sign in to your workspace if you aren’t already. Pick the workspace, approve the permissions. Slack hands you back to the dashboard.

3

Pick a channel

The card pulls the channel list from your workspace into a dropdown. Pick where alerts should land — #support, #leads, #founders, whatever your team’s convention is.

4

Toggle events

Three default-on events — lead captured, escalation, fallback — are the right starting point for almost every team. Toggle the others as you need them.

5

Send test message

A test alert lands in your selected channel in seconds. If anything’s misconfigured, the card shows you the error.

The whole connection flow, including channel selection and the test message, takes about two minutes from clicking Connect to seeing the first real alert in your channel.

Per-bot routing. If you want different bots to alert different channels — your sales bot pings #leads and your support bot pings #support— connect Slack on each bot independently. Each bot owns its own channel and event preferences. Same Slack workspace, no collision.

04

Real teams using it

What this actually looks like in production, in the kinds of teams who run NebulaHex day to day:

🚀

Founder watching first widget go live

You launched the website widget last Tuesday. Instead of refreshing the dashboard every five minutes, you pipe lead captures and escalations into #founders. When someone asks for a demo, you see it at your desk. When the bot hits a hard question, you decide whether to jump in personally or let the team take it.

Bot becomes a tool you actually use
🛠️

Small support team running Slack-first

Your team of four already does everything in Slack. You have #support for triage, #wins for sales, #oncall for first responders. Adding NebulaHex’s alerts means your bot fits into the system you already have — lead captures land in #wins, escalations land in #support, fallbacks land in #bot-feedback for the knowledge owner.

No new dashboard to babysit
🛒

E-commerce catching abandoned questions

Your store sells supplements. The bot answers most product questions on its own. But when someone asks about a medication interaction the bot rightly escalates, your small ops team gets pinged within seconds. One of them takes the WhatsApp or widget thread over before the customer has time to close the tab.

Response time becomes revenue
📊

Consultancy using fallbacks as content strategy

You run an agency. Your help docs are decent but not exhaustive. Fallback alerts pipe into a channel your content lead checks weekly. Each one signals a real prospect asked a real question your knowledge base couldn’t answer well. After two weeks, that channel becomes a backlog of FAQ entries, blog posts, and case studies.

Bot becomes market research

05

Why native, not Zapier

You could glue a Slack notification to your bot using a webhook and a third-party automation platform — Zapier, Make, or any of the others. We support that pattern too; webhooks are first-class in NebulaHex. But the native integration exists because the Zapier path has friction the native one doesn’t.

Via Zapier / webhooks

  • Build a Zap connecting webhook to Slack
  • Map payload fields to message format
  • Manage monthly automation quota
  • Pay for second tool to keep it running
  • Maintain when payloads change

Native integration

  • Click Connect Slack once
  • Pick a channel from dropdown
  • Tick the events you want
  • Send test message — done
  • Updates ship when we update them

The native integration gives you the right defaults. You don’t have to think about which events to subscribe to or how to filter them. The three default-on events (lead, escalation, fallback) are the three almost every team wants from day one. The two default-off events (every message, test results) are off because we know they create noise for most users.

If you ever need more flexibility — custom formatting, branching on event content, fan-out to other tools — the webhook and Zapier paths are still there, sitting alongside the native integration. Most teams never need them.

06

Trust & security

The Slack connection uses standard OAuth 2.0. You authenticate against Slack directly — NebulaHex never sees your password, only the access token Slack issues after you grant the permissions. That token is encrypted at rest, and integration permissions are scoped to what’s needed to post messages and read the channel list. We don’t read your team’s conversations, we don’t index your channels, and we don’t store the content of messages your team sends in Slack.

Outbound calls to Slack happen over TLS. The token is rotated when you reconnect. If you disconnect the integration from the dashboard, the token is revoked on our side immediately, and you can also revoke from inside Slack’s app management screen at any time — subsequent calls will fail cleanly.

Access to integration settings is controlled by the same role-based access control that governs the rest of your NebulaHex workspace. Owners and Admins can connect, configure, and disconnect. Editors can see the configuration but not change it. Viewers see only what’s already configured.

We don’t use your conversation data to train AI models. The knowledge sources you upload power your bot’s answers, not anyone else’s. See our privacy policy and terms for the full posture.

07

Connect Slack to your bot in minutes

If you already have a NebulaHex bot, connecting Slack is the next thing to do. Open your bot, go to the Integrations tab, click Connect Slack, sign in, pick a channel, toggle on the three defaults. Send a test message. Two minutes later you’re getting alerts in the channel your team already lives in.

Your bot does the work

Slack tells you when the work needs you. The right division of labor between automation and people that makes a bot worth running.