How to add an AI chatbot to Slack

Your team already lives in Slack. When someone needs the VPN reset steps, the PTO policy, or a ship date, they drop the question in a channel and wait. The same answers get retyped every week, and the one person who knows is heads-down.

Supportweave turns Slack into one clean intake. By default, every question in a channel or DM is answered by AI — the same knowledge-grounded pipeline that runs on your website widget — and the grounded reply goes straight back out in Slack, pulled from your own docs. Prefer a person in the loop? Flip the AI replies toggle off and that channel goes human-first: the question lands in a shared inbox flagged for a human, with a desktop alert on arrival and an AI copilot that drafts the answer for your teammate to review and send. It works the same for a customer community as it does for an internal IT or HR desk.

What you'll need

• A Supportweave workspace (set up during your demo)
• A trained chatbot whose knowledge the AI answers from
• A Slack workspace where you can create and install apps (admin approval may be required)

Step 1: Train your agent

Before the AI can answer, it needs your knowledge. Crawl your site, upload docs, or add FAQs so it has something to pull from. Our training guide walks through it step by step.

Step 2: Create a Slack app

Go to api.slack.com/apps and click Create New App. Choose From scratch, give it a name, and pick the workspace it should live in.

Step 3: Add bot token scopes

Open OAuth & Permissions. Under Bot Token Scopes, add chat:write so replies can go back out through the bot. To pick up direct messages, add im:history. To pick up messages in channels the bot is invited to, add channels:history. Add whichever ones match where you want to field questions.

Step 4: Install the app and copy your credentials

Still in OAuth & Permissions, click Install to Workspace and approve the scopes. Once it installs, copy the Bot User OAuth Token — it starts with xoxb-. Then open Basic Information → App Credentials and copy the Signing Secret. You'll paste both into Supportweave next.

Step 5: Connect the channel in Supportweave

In the Supportweave dashboard, open your chatbot and go to its Channels tab, pick Slack, and paste in the Bot User OAuth Token and the Signing Secret you just copied. Connect the channel, and Supportweave hands you a webhook URL — copy it. Channels are set per chatbot, so Slack routes to the agent whose knowledge you just trained.

Step 6: Turn on Event Subscriptions in Slack

Back in your Slack app, open Event Subscriptions and turn it on. Paste the Supportweave webhook URL into the Request URL field. Slack sends a verification challenge, and Supportweave answers it automatically, so the field verifies on its own — and every event after that arrives signed, with the signature verified on the way in. Then, under Subscribe to bot events, add message.im for DMs and message.channels for channels the bot is in.

Step 7: Invite the bot and ask a real question

Invite the bot to any channel it should watch by typing /invite @YourBot in that channel — direct messages work as soon as the app is installed, no invite needed. Then send the bot a DM or post in a channel it's in.

Within moments the AI answers right in Slack — a grounded reply pulled from your knowledge base, citing its sources — and the whole exchange shows up in your Supportweave inbox as a conversation tagged Slack, so your team can follow along. If someone asks for a person (or the workspace is out of credits), the thread flips to needs a human: it enters the wait queue with its true arrival time, fires a desktop alert if you've turned them on, and the customer gets a quick note that a teammate is on the way.

Prefer a person on every message? Flip the AI replies toggle off for this channel on the Channels tab, and Slack goes human-first — the question lands flagged for a human, and the AI copilot drafts a suggested reply for your teammate to review, edit, and send back out in Slack on the same thread.

Take it further

• Point the AI at your internal IT and HR docs so every answer is grounded in the right policy, not a guess
• Every message is auto-tagged with intent, sentiment, and a lead score, and the copilot writes a handoff summary for whoever picks up a human-first thread
• Repeat askers merge into one contact timeline, so you see their whole history across every channel
• Add WhatsApp, SMS, Instagram, or email from the same agent, same knowledge, no retraining — the chat channels answer with AI by default, email stays human-first, all in one inbox
• The web widget and a voice phone line answer on their own too — instant, grounded replies with no human in the loop
• White-label it so replies read as your brand, not ours

Slack is where your team and community already talk, so let the AI answer every channel and DM by default, grounded in your docs — and flip to a human, backed by an AI copilot, whenever you want. Either way it all lands in one inbox. See the Slack channel page, or book a demo to see it in a workspace.