Uniki User Guide

Appearance

3.5: Slack Integration ​

Imagine your team has a question about company policies, product information, or support procedures. Instead of interrupting someone or digging through documents, they could just type:

@SupportBot What's our return policy?

And get an instant, accurate answer—right in Slack, where they're already working.

That's what we're building today: A Slack bot powered by AI that becomes your team's instant knowledge assistant.

Here's what makes this powerful:

  • Your team stays in Slack (no context switching)
  • AI searches your knowledge base instantly
  • Answers are consistent and always available
  • Works 24/7, never takes a break
  • Learns from your company's specific information

Let's build this together!

3.5.1: What You'll Build ​

Here's your Slack bot in action:

#
support-team
Bot Active
4 members
SC
Sarah Chen10:23 AM
@SupportBot What's our return policy?
🤖
SupportBot APP 10:23 AM
Our return policy allows you to return most items within 30 days of purchase for a full refund. Items must be in original condition with tags attached. Would you like me to help you start a return?
MJ
Mike Johnson10:25 AM
@SupportBot Can you check the status of order #12345?
🤖
SupportBot APP 10:25 AM
Order #12345 was shipped yesterday via UPS and is currently in transit. Expected delivery: Friday, Nov 24th. Tracking number: 1Z999AA1234567890
Message #support-team...
💡 Mention @SupportBot to ask questions

The workflow:

  1. Team member mentions bot → "@SupportBot What's our return policy?"
  2. Bot appears in your taible → New row created with the question
  3. AI generates answer → Searches knowledge base, creates response
  4. Response sent to Slack → Appears in the same channel, same thread
  5. Team member gets instant help → Without leaving Slack!

💡 Why This Is a Game-Changer

Traditional way: Team member asks in Slack → Someone has to stop working → Look up the answer → Type response

  • Takes 5-15 minutes
  • Interrupts someone's focus
  • Answer quality varies by who responds

With your AI Slack bot:

  • Instant responses (5-10 seconds)
  • Nobody gets interrupted
  • Consistent, accurate answers every time
  • Works 24/7 including weekends

Result: Your team stays productive, information flows freely, and everyone has access to collective knowledge!

3.5.2: Why Slack Is Different ​

Before we build, let's understand why Slack integration is unique.

💡 💡 Slack vs Email vs Chat Widget

Email (Section 3.4):

  • Customer-facing, formal
  • Needs approval before sending
  • One-to-one communication

Chat Widget (Section 3.2):

  • Customer-facing, public website
  • Anonymous visitors
  • Support-focused

Slack:

  • Internal team communication
  • Known team members
  • Fast, informal, conversational
  • Context-rich (channels, threads, mentions)

Key difference: Slack is for your TEAM, not customers. This means:

  • Faster responses (no approval needed)
  • Access to internal knowledge
  • Can reference company-specific context
  • More casual, friendly tone

3.5.3: Step 1 - Connect Your Slack Workspace ​

First, we need to authorize Taibles to access your Slack workspace.

What we're doing: Creating a secure connection between Taibles and Slack so your bot can read mentions and send responses.

Connect Slack Workspace

Authorize Taibles to access your Slack workspace

Taibles will be able to:

  • Read messages when your bot is mentioned
  • Send messages and replies to channels
  • Access basic workspace information

Select channels your bot can access:

Your bot will only respond when mentioned (@SupportBot) in these channels

Secure Connection: Your Slack credentials are encrypted and stored securely. You can revoke access anytime from your Slack workspace settings.

Let's do it:

  1. Open your taible (or create a new one called "Team Support Bot")

  2. Click "Add Trigger" button in the toolbar

    • Look for the purple button with a lightning bolt icon

  3. Find "Slack" in the triggers list

    • Search for "Slack" or browse the "Team Communication" category

  4. Click "Slack Message" trigger

  5. The Slack authorization popup appears

    • Click "Authorize Slack Workspace"
    • A new window opens asking you to sign in to Slack

  6. Select your workspace

    • If you're in multiple Slack workspaces, choose the right one
    • This is where your bot will live

  7. Review permissions

    • The bot will be able to:
      • ✅ Read messages when mentioned
      • ✅ Send messages and replies
      • ✅ Access basic workspace info
    • These are the minimum permissions needed

  8. Select channels

    • Choose which channels your bot can access
    • Recommended for team bots:
      • #support-team - Customer support questions
      • #general - General company questions
      • #sales - Sales-related questions
    • You can always add more channels later

  9. Click "Allow" to authorize

  10. You're connected! ✨

    • You'll see a success message
    • Your Slack workspace is now connected

💡 🔒 Security & Privacy

What the bot CAN do:

  • Read messages only when @mentioned (it can't spy on conversations!)
  • Send responses in channels you selected
  • Access basic public workspace info (channel names, user names)

What the bot CANNOT do:

  • Read private messages or DMs (unless invited)
  • Read channels you didn't select
  • Delete or edit others' messages
  • Access sensitive workspace settings

Revoking access: You or any Slack admin can revoke the bot's access anytime from Slack's App Management settings.

3.5.4: Step 2 - Configure the Slack Trigger ​

Now let's configure exactly how your bot receives messages.

Configure Slack Trigger

Step 2 of 3

When someone mentions your bot in Slack (e.g., @SupportBot), a new row will be created in your taible with the message details.

Bot will create rows when mentioned in these channels

Data Captured in Each Row:

Message: The full text of the message
User: Who mentioned the bot (name and email)
Channel: Which channel the message came from
Thread ID: For replying in the same thread
Timestamp: When the message was sent
Example Row Created:
message: "What's our return policy?"
user: "Sarah Chen (sarah@company.com)"
channel: "#support-team"
timestamp: "2024-11-19 10:23:45"

Let's set it up:

  1. The trigger configuration panel opens automatically

    • If not, click the ⚙️ settings icon on your Slack trigger

  2. Verify your workspace is selected

    • Should show: "Your Company Workspace" (or your workspace name)

  3. Review monitored channels

    • Checkboxes show which channels the bot watches
    • Selected channels are where the bot will respond when mentioned
    • Recommendation: Start with 1-2 channels, expand later

  4. Understand what gets captured

    • When someone mentions @YourBot in a monitored channel:
      • Message text → The full question
      • User info → Name and email of who asked
      • Channel → Which channel it came from
      • Thread ID → So you can reply in the right place
      • Timestamp → When the message was sent

  5. Click "Save Trigger"

What happens now:

Every time someone mentions your bot in Slack (e.g., "@SupportBot how do I reset my password?"), the trigger fires and creates a new row in your taible with all the message details.

💡 💡 Pro Tip: Bot Naming

Choose a clear bot name that makes its purpose obvious:

  • ✅ @SupportBot - For support questions
  • ✅ @PolicyBot - For company policies
  • ✅ @HRHelper - For HR questions
  • ✅ @KnowledgeBot - For general knowledge

Your team will mention it naturally when they need help!

3.5.5: Step 3 - Set Up the AI Response Column ​

Now let's add AI to actually answer the questions!

This is the same AI Assistant column we've used before, but we'll customize it for team communication.

Let's create it:

  1. Click "+ Add Column"

  2. Choose "AI Assistant" type

    • Found in "Talk to Customers" category

  3. Name it: ai_response

  4. Label: "AI Answer"

  5. Add dependency:

    • The AI needs to wait for the Slack message to arrive first
    • Click "Add Dependency"
    • Select the trigger's message field
    • This ensures AI has the question before answering

  6. Write instructions for your team bot:

You are a helpful team assistant for [Your Company Name].

When team members ask questions, provide clear, accurate answers based on our knowledge base.

Guidelines:
- Search our internal knowledge base for accurate information
- Be friendly and conversational (we're talking to teammates!)
- If you don't know something, say so honestly
- Keep answers concise but complete
- Reference specific documents or policies when relevant
- If the question needs a human expert, suggest who to ask

Remember: You're helping colleagues, not customers, so you can be more casual and direct.

  1. Enable "Search Knowledge Base"

    • Toggle this ON
    • This lets AI search your company docs, policies, procedures
    • (Make sure you've uploaded your knowledge base in Settings → Knowledge)

  2. Choose AI model (optional)

    • Default is usually fine (GPT-4)
    • For faster responses, consider GPT-3.5
    • For more thoughtful analysis, stick with GPT-4

  3. Click "Save"

💡 Customizing for Your Team

Adjust the tone for your company culture:

Startup/Casual:

Hey! I'm here to help with quick answers. I'll search our docs and give you the scoop. If I'm not sure about something, I'll point you to someone who knows!

Corporate/Professional:

I'm your team's knowledge assistant. I provide accurate information from our official documentation and policies. For matters requiring management approval, I'll direct you to the appropriate department.

Technical Team:

I'm your technical knowledge bot. I can help with documentation, procedures, and technical references. For complex debugging, I'll point you to subject matter experts.

Match your team's communication style!

3.5.6: Step 4 - Add the Slack Output Column ​

Now for the final piece: sending the response back to Slack!

Add Slack Output Column

Step 3 of 3

This column will send your AI-generated response back to Slack in the same channel where the user asked their question.

This will send the content from the selected column

Reply Settings:

When to Send:

This column will run automatically after the AI Response column completes.
💡 The message will be sent as soon as your AI generates a response—usually within 5-10 seconds.
How it works:
1User mentions @SupportBot in Slack
2Trigger creates new row with message
3AI Response column generates answer
4This column sends response back to Slack
5User sees answer instantly in Slack!

Let's set it up:

  1. Click "+ Add Column"

  2. Choose "Send to Slack" type

    • Found in "Team Communication" category

  3. Name it: send_to_slack

  4. Label: "Send Slack Response"

  5. Configure Slack settings:

    Account: Your workspace (already connected)

    Message content: Click "AI Answer" button

    • This inserts the AI-generated response

    Reply settings (automatically configured):

    • ✅ Reply in same channel (where the question was asked)
    • ✅ Reply in thread (keeps conversations organized)

  6. Set when to run:

    • Add dependency: "AI Answer"
    • This means the response sends automatically after AI finishes
    • No approval needed for internal team communication!

  7. Click "Save"

💡 💡 Why No Approval for Slack?

Remember the difference:

Email to customers: Needs approval

  • Public-facing
  • Permanent record
  • Formal communication
  • Represents your brand

Slack to team: Automatic (no approval)

  • Internal communication
  • Fast, conversational
  • Team knows it's a bot
  • Can be quickly corrected if needed

If the bot makes a mistake in Slack:

  • Just post a correction
  • Team understands it's AI
  • Fast iteration improves accuracy

This speed is what makes Slack bots so valuable!

3.5.7: Step 5 - Test Your Slack Bot ​

Time to see it in action!

Let's test:

  1. Open Slack (your connected workspace)

  2. Go to one of your monitored channels

    • For example: #support-team or #general

  3. Mention your bot with a question:

    • Type: @SupportBot What's our return policy?
    • Or: @SupportBot How do I request time off?
    • Or: @SupportBot What are our business hours?

  4. Press Enter to send

  5. Watch what happens:

    • A moment of suspense... (about 5-10 seconds)
    • Your bot replies with an answer!
    • The response appears in the same thread

  6. Check your taible (optional):

    • Refresh your taible view
    • You'll see a new row with:
      • The user's message
      • AI's response
      • Slack status: "Sent ✓"

Slack Support Bot

3 conversations

User Channel Message AI Response Slack Status Time
SC
Sarah Chen
#support-team
What's our return policy?
Our return policy allows you to return most items within 30 days...
Sent10:23 AM
MJ
Mike Johnson
#support-team
Can you check the status of order #12345?
Order #12345 was shipped yesterday via UPS...
Sent10:25 AM
ED
Emily Davis
#general
What are your business hours?
We're open Monday-Friday 9am-6pm EST...
Sending...10:27 AM
💡 Each row represents one Slack mentionResponses send automatically when AI completes

Try more questions:

  • Ask about company policies
  • Request information from your knowledge base
  • Test with different team members
  • Try in different channels

💡 🎉 You Built a Team Knowledge Bot!

Congratulations! Your Slack bot is live and helping your team!

Your bot now:

  • ✅ Listens for @mentions in Slack
  • ✅ Creates rows automatically
  • ✅ Searches your knowledge base
  • ✅ Generates accurate answers with AI
  • ✅ Responds instantly in Slack
  • ✅ Works 24/7 for your entire team

The impact:

  • Teammates get instant answers
  • Nobody needs to interrupt their work
  • Knowledge flows freely
  • New team members can self-serve
  • Your subject matter experts stay focused

You just made your whole team more productive!

3.5.8: Customizing Your Slack Bot ​

Now that the basics work, let's make it even better!

1. Add Multiple Channel Support ​

Want the bot in more channels?

In your Slack trigger settings:

  • Click the ⚙️ settings icon
  • Add more channels to the monitored list
  • Common additions:
    • #engineering - Technical questions
    • #sales - Sales process questions
    • #onboarding - New employee questions
    • #product - Product information

2. Channel-Specific Responses ​

Want different behavior by channel?

Add a "Channel Router" column:

  1. Add a new AI Assistant column
  2. Use this logic:
    If the question came from #engineering:
  - Focus on technical accuracy
  - Reference code documentation
  - Use technical terminology

If from #sales:
  - Focus on product benefits
  - Include pricing information
  - Use customer-facing language

3. Add Quick Response Templates ​

For common questions, add faster responses:

Create a "Quick Check" column before AI:

  • Type: Custom Code or Rules
  • Check if question matches common patterns:
    • "business hours" → "We're open Monday-Friday, 9am-6pm EST"
    • "vacation policy" → Link to HR policy doc
    • "password reset" → Link to IT self-service
  • If no match, let AI handle it

This makes common questions instant!

4. Escalation to Humans ​

For questions AI can't handle:

In your AI instructions, add:

If the question is about:
- Personal HR matters → "@hr-team can help with that!"
- Financial approvals → "@finance-team handles those"
- Technical emergencies → "@eng-oncall for urgent issues"

Always suggest the right person or team for complex matters.

5. Add Feedback Collection ​

Let your team rate responses:

In your Slack output column:

  • Enable "Add reaction buttons"
  • Adds 👍 and 👎 buttons to bot responses
  • Track in a "Feedback" column
  • Review periodically to improve instructions

3.5.9: Advanced: Multi-Bot Strategy ​

As your needs grow, consider multiple specialized bots:

Strategy 1: One Bot, Many Purposes ​

Single bot handles all questions

  • ✅ Simple for team (one bot to remember)
  • ✅ Easier to maintain
  • ❌ Less specialized responses
  • ❌ Harder to customize per use case

Strategy 2: Specialized Bots ​

Multiple bots for different purposes

  • ✅ Highly specialized and accurate
  • ✅ Clear purpose for each bot
  • ❌ Team needs to remember which bot for what
  • ❌ More setup and maintenance

Example multi-bot setup:

  • @SupportBot - Customer support procedures
  • @HRBot - HR policies and benefits
  • @ProductBot - Product information
  • @ITBot - Technical help and troubleshooting

Choose based on:

  • Team size (bigger teams = more specialization helpful)
  • Knowledge base complexity
  • Department separation
  • Response quality needed per domain

3.5.10: Best Practices ​

🎯 Keep responses concise:

  • Slack is for quick answers
  • Aim for 2-3 sentences max
  • Provide links for details
  • Use bullet points for lists

🚀 Monitor and improve:

  • Check your taible weekly
  • Look for questions the bot struggled with
  • Update knowledge base with missing info
  • Refine AI instructions based on patterns

📊 Track usage:

  • Add a dashboard (we'll cover in Section 7.3)
  • Track: Questions per day, channels used, response quality
  • Celebrate when adoption grows!

🔒 Maintain security:

  • Don't put sensitive data in public channels
  • Review bot permissions quarterly
  • Keep knowledge base current
  • Monitor for inappropriate questions

👥 Train your team:

  • Announce the bot in team meeting
  • Share example questions
  • Encourage feedback
  • Create a "bot tips" document

3.5.11: Troubleshooting ​

"Bot isn't responding in Slack"

  • Verify trigger is enabled (check for ⚡ icon)
  • Confirm you @mentioned the bot (not just typed its name)
  • Check bot has access to that channel
  • Look for errors in the taible rows

"Responses are too slow"

  • Check if knowledge base is too large (split it up)
  • Consider using a faster AI model (GPT-3.5 vs GPT-4)
  • Remove unnecessary dependencies
  • Check for rate limiting

"Bot gives wrong answers"

  • Update your knowledge base with correct info
  • Improve AI instructions with specific examples
  • Add fact-checking to instructions
  • Consider adding human review for critical topics

"Team isn't using the bot"

  • Make the bot's purpose clearer (better name, description)
  • Share success stories in team meetings
  • Add bot mentions to onboarding docs
  • Request feedback: What would make it more useful?

"Bot responds in wrong channel"

  • Check trigger channel configuration
  • Verify Slack account permissions
  • Test in a private channel first
  • Review Slack output column settings

What You've Learned ​

Incredible work! You've built a production-ready Slack bot for your team.

You now know how to:

  • ✅ Connect Taibles to your Slack workspace
  • ✅ Configure triggers to listen for @mentions
  • ✅ Set up AI to answer team questions
  • ✅ Send responses back to Slack automatically
  • ✅ Customize the bot for your team's needs
  • ✅ Monitor and improve bot performance

The power of this:

  • Your team has instant access to collective knowledge
  • Nobody needs to interrupt their flow to ask questions
  • New team members can self-serve
  • Information is consistent and always available
  • Your experts can focus on complex work

Real-world impact:

  • Support teams: Answer internal procedures instantly
  • Sales teams: Product info always at hand
  • Engineering: Quick access to technical docs
  • HR: Policies and benefits on demand
  • Everyone: Less time searching, more time doing

Next up: In section 3.6, we'll connect your agent to Telegram so you can reach users on another popular messaging platform!

Previous: 3.4 Email IntegrationNext: 3.6 Telegram Integration

Built with VitePress

Copyright © 2025 Uniki