AI Model Wiki

    Skip to main content
    18 min
    Lesson 2 of 4

    AI Workflows and Custom Agents

    From the course Slack AI Agents

    Build AI-powered workflows and custom agents in Slack for automated team assistance.

    Workflow Builder with AI

    AI Steps in Workflows

    • Generate Text: AI creates content within workflow steps
    • Summarize: Condense information from multiple sources
    • Classify: Route messages based on AI analysis
    • Extract: Pull specific data from unstructured messages

    Example: Support Ticket Workflow

    Trigger: Message posted in #support-requests
    Steps:
    1. AI: Classify issue type (billing/technical/feature)
    2. AI: Assess priority (P1/P2/P3)
    3. AI: Generate acknowledgment message
    4. Route to appropriate channel based on classification
    5. Create ticket in external system via webhook
    6. Notify assigned team member
    

    Example: Meeting Follow-up

    Trigger: Message posted in #meeting-notes
    Steps:
    1. AI: Extract action items from meeting notes
    2. AI: Identify assignees from context
    3. Create tasks in project management tool
    4. Post summary to relevant project channels
    5. Schedule follow-up reminder in 1 week
    

    Custom AI Agents (Slack Agents)

    Building an Agent

    • Define the agent's purpose and scope
    • Configure knowledge sources (channels, files, external data)
    • Set up conversation patterns and responses
    • Deploy to specific channels or as DM bot

    Agent Use Cases

    • IT Help Desk: Answers common IT questions from documentation
    • HR Assistant: Employee handbook and policy queries
    • Sales Support: Product information and pricing lookups
    • Onboarding Guide: New employee questions and orientation

    Third-Party AI Integrations

    • ChatGPT app for Slack
    • Claude for Slack
    • Custom bots via Slack API + AI services
    • Zapier/Make AI automations triggered from Slack

    Key Takeaways

    • •Workflow Builder AI steps handle text generation, classification, and data extraction
    • •Support workflows automatically categorize, prioritize, and route incoming requests
    • •Custom AI agents answer domain-specific questions from configured knowledge sources
    • •Third-party AI integrations extend capabilities with ChatGPT, Claude, and custom bots
    0 of 4 lessons marked complete

    Progress is kept in this browser only.