Bringing agents to Slack transforms how your team learns and adopts AI automation:
Shared Learning
Visibility by default: Every interaction in a channel becomes a learning opportunity for the team. No more siloed knowledge—everyone sees how to use agents effectively.
Natural Integration
Where work happens: Teams already use Slack for communication. Agents integrate seamlessly into existing workflows without requiring new tools.
Collaborative Usage
Team-wide access: Instead of one person running automations, entire teams can leverage the same agent with consistent results.
Instant Adoption
Zero learning curve: If your team knows how to @mention someone in Slack, they know how to use an agent. No training required.
The Learning Effect: When someone asks an agent a question in a channel, everyone sees the interaction. This passive learning accelerates team-wide adoption faster than any training session could.
If you haven’t already connected Gumloop to your Slack workspace, you’ll need to authenticate first.Visit the personal credentials page and authenticate with Slack to connect the Gumloop app to your workspace.
Click the Add Credential button for Slack and follow the OAuth flow to authorize Gumloop in your workspace.
2
Add the Gumloop bot to Your Channel
In the Slack channel where you want your agent you would need to add the Gumloop bot. To add the bot, you have two options:
Type /invite @Gumloop in the channel, OR
Click the channel name → Add integrations/Add app → Search for “Gumloop”
This gives Gumloop permission to read and respond in the channel.
Agents can be added to public or private channels, but not to direct messages (DMs).
3
Get Your Agent's Slack Command
In Gumloop, navigate to the Agents page and click on the agent you want to add to Slack.
To interact with your agent, @mention it in a top-level message:
Copy
Ask AI
@Gumloop [your question or task]
Important: You must @Gumloop the agent for all top-level messages. The agent will not respond to messages that don’t include an @Gumloop.
The agent will:
Start a thread to keep conversations organized
Process your request using its configured tools and workflows
Respond with results and reasoning
Continue the conversation within the thread
Need custom branding? If you want your agent to appear with a custom name and avatar instead of “@Gumloop”, you can connect your own Slack app. See Custom Slack App Integration for setup instructions.
When to respond to a message in a thread.This setting determines whether your agent responds to every message in a thread or only when explicitly mentioned.
On All Messages
Only on Mentions (Recommended)
How it works: Agent responds to every message in the thread, even without @Gumloop.Best for:
Active collaboration where the agent should participate in all discussion
When multiple team members are asking questions
Continuous back-and-forth conversations
Support channels where the agent acts as a primary responder
Example: After starting a thread with @Gumloop analyze this customer, any follow-up like “What about their purchase history?” will automatically trigger the agent.
How it works: Agent only responds when explicitly @mentioned in thread replies.Best for:
Allowing team discussions without agent interruption
More controlled agent participation
Channels where both human and agent input are needed selectively
Letting teams discuss agent responses before requesting follow-ups
Example: After @Gumloop analyze this customer, you can discuss the results with your team in the thread. The agent only responds again when you @Gumloop with a follow-up question.
Use this mode when your team needs to discuss the agent’s responses without triggering new responses. Simply reply in the thread without @mentioning the agent to have private team discussions.
See the agent’s thought process as it responds to your messages.By default, this setting is disabled (recommended for cleaner threads).
When to Enable Stream Reasoning
Disabled (Default - Recommended):
Agent shows only the final response
Threads remain clean and focused
Best for most production use cases
Team members see polished answers without technical details
Enabled:
Shows real-time reasoning as the agent processes your request
Displays each tool call, decision, and intermediate step
Useful for debugging agent behavior
Helps understand how the agent arrived at conclusions
Keep this disabled unless you need debugging visibility. Most users prefer cleaner threads without the technical reasoning details. Enable only when you want complete transparency into the agent’s decision-making process.
Example of streamed reasoning:
Copy
Ask AI
🤔 Analyzing the customer request...🔍 Searching CRM for customer ID #12345...✅ Found customer record🔍 Checking purchase history in database...📊 Analyzing 15 transactions from past 6 months...💡 Generating summary...
versus just the final response:
Copy
Ask AI
Customer #12345 has made 15 purchases totaling $3,450 over the past 6 months. Their average order value is $230.
Hide the view full flow results button while chatting.By default, this setting is disabled, showing workflow run details in threads.
Understanding Flow Run Results
When your agent uses workflows (flows) to complete tasks, Gumloop can show a “View full flow results” button in the thread along with workflow run details.Disabled (Default):
Shows “View full flow results” button in thread
Displays workflow run metadata
Provides transparency into which workflows were triggered
Useful for debugging and understanding automation flow
Enabled (Cleaner Threads):
Hides the workflow button and run details
Agent simply returns with its response
Creates cleaner, more conversational threads
Recommended when team doesn’t need workflow visibility
Enable this setting for cleaner threads. Most end users don’t need to see workflow execution details—they just want the agent’s answer. This makes conversations feel more natural and less technical.
With workflow details shown (disabled):
Copy
Ask AI
Agent: I've processed the customer data.[View full flow results →]Workflow: Customer Analysis PipelineRun ID: abc123xyzDuration: 3.2sStatus: ✅ CompletedResult: Customer is eligible for discount tier 2
With workflow details hidden (enabled):
Copy
Ask AI
Agent: I've processed the customer data.Result: Customer is eligible for discount tier 2
These settings can be changed anytime in your agent’s Slack preferences. Changes take effect immediately for new conversations.
When your agent generates images using the Image Generation tool, both the image file and a link to the image are sent directly in the Slack thread. This makes it easy to view the generated image inline and also access or share the image URL.To enable image generation for your agent, add the Image Generation tool in the agent builder under Tools → Add Image Generation Tool:
Image generation costs 30 credits per image regardless of the model used (DALL-E 3, Gemini 2.5 Flash, Gemini 3 Pro, or GPT-Image-1). This is charged in addition to the agent’s normal model costs.
When you interact with an agent in Slack, the agent uses your personal default credentials (unless workspace credentials are configured)—not the agent creator’s credentials. This ensures proper access control and data privacy.
For Existing Gumloop Users
If your Slack email matches your Gumloop account:
The agent automatically uses your personal default credentials
You have immediate access to all tools and workflows you’re authorized to use
No additional setup required
Example: If the agent uses Gmail and Google Calendar, it will access your personal Gmail and Calendar using your authenticated credentials.
For Non-Gumloop Users
If you’re not yet a Gumloop user, you’ll see a signup prompt the first time you try to use an agent:
What happens:
Agent responds with a message asking you to sign up
Click the link to create your Gumloop account using your Slack email
After signing up, authenticate with the services the agent needs
Return to Slack and @Gumloop the agent again
You must use the same email address for Slack and Gumloop for the integration to work properly.
Missing Credentials
If the agent needs a credential you haven’t set up yet:
The agent will notify you about the missing authentication
Example: If an agent uses Google Calendar but you haven’t authenticated, you’ll get a message like:
Copy
Ask AI
⚠️ I need access to Google Calendar to complete this task. Please authenticate at: https://www.gumloop.com/settings/profile/credentials
Personal vs Workspace Agents
Recommendation: Always create agents in your personal workspace unless you need shared workspace credentials or team collaboration features.
Personal Agents (created in personal workspace):
Anyone in the Slack channel can use the agent
Each user’s request runs on their own personal credentials
Non-Gumloop users will be prompted to sign up
Best for most use cases
Workspace Agents (created in shared workspace):
Access control: Only members of that specific Gumloop workspace can use the agent
Credential behavior: If an MCP integration or credential is set to use “workspace default,” the workspace credentials are used instead of personal credentials
Non-workspace members will receive an access denied message
If you want team members to use your agent without requiring them to set up their own credentials, you can configure the agent to use workspace credentials. This is particularly useful when you want to share an agent with your team and have everyone use the same authenticated accounts.
Team or Enterprise plan required: Workspaces are available on Team and Enterprise plans only.
1
Create or Select a Workspace
First, ensure you have a workspace set up. You can check your existing workspaces or create a new one from the hub page by clicking the workspace dropdown and selecting ”+ Add a workspace”.
Alternatively, you can use your organization’s default workspace, which automatically includes all organization members. Find it in your organization settings under the General tab.
2
Move the Agent to the Workspace
From the hub page, find your agent and click the three-dot menu. Select “Move to Workspace” and choose the workspace you want to use.
3
Add Workspace Credentials
Open the workspace settings by navigating to the workspace and clicking on “Workspace Credentials” in the left sidebar. Add the credentials that your agent needs (e.g., Gmail, Google Drive, Salesforce).
Members of your workspace will be able to use the integrations you link here through flows and agents.
4
Configure Agent to Use Workspace Credentials
In your agent’s settings, switch each integration from “Use Personal Default” to “Use Workspace Default”. Click the three-dot menu next to each integration and select the workspace credential option.
Important: When an agent is in a workspace, users invoking the agent must also be members of that workspace. Otherwise, they will not be able to use the workspace credentials or invoke the agent. To avoid manually adding users, use your organization’s default workspace which automatically includes all members.
When you use an agent with personal credentials, only your authenticated credentials are used. Other team members cannot access your personal data through the agent.
Controlled Access
Admin security controls and user roles in Gumloop apply to agents just like they do to workflows. You can only access what you’re authorized to access.
Deploy agents strategically to appropriate channels:Supported channel types:
✅ Public channels
✅ Private channels
❌ Direct messages (DMs) - Not supported
Example deployments:
#support: Customer service agents with ticket triage capabilities
#sales: Lead research and enrichment agents
#marketing: Campaign planning and content strategy agents
#data: Data analysis and reporting agents
Match agent capabilities to channel purpose for maximum adoption.
Set Channel Expectations
When adding an agent to a channel, post a message explaining:
Copy
Ask AI
📢 New Agent Alert!I've added our "Support Bot" agent to this channel.What it can do:• Look up customer tickets in Zendesk• Check user eligibility for discounts• Pull order history from SalesforceHow to use it:@Gumloop [your question]Example: "@Gumloop Check if ticket #12345 is eligible for a discount"💡 Remember: You must @Gumloop the agent in all top-level messages!
This reduces friction and encourages proper usage.
Use Descriptive Agent Names
When your team @mentions “@Gumloop”, they’re interacting with whichever agent is active in that channel.Use clear agent names in your Gumloop workspace:
✅ “Support Ticket Assistant”
✅ “Sales Lead Researcher”
✅ “Marketing Campaign Planner”
❌ “Agent 1”
❌ “My Bot”
Team members can check which agent is active with /gummie active.
Configure Thread Response Mode Appropriately
Choose the right thread response mode in your agent’s Slack settings:
“On all messages”: Best for agents that should participate in every reply
Support bots that need to track all conversation
FAQ agents that should answer follow-up questions
Active collaboration scenarios
“Only on mentions” (Recommended): Best for most use cases
Research assistants used occasionally
Agents where teams need to discuss responses
Controlled agent participation
This prevents unwanted interruptions and makes agent behavior more predictable.
Monitor Usage Patterns
Watch how your team uses agents in Slack:
Which questions get asked most frequently?
Where do agents struggle or give poor responses?
What workflows are called most often?
Use these insights to:
Refine agent instructions
Add new tools and workflows
Improve response accuracy
Create training materials for your team
You can view all agent conversations in the agent’s settings page in Gumloop.
Ensure Credentials Are Set Up
Before deploying an agent to a team channel:
Test the agent with your own credentials
Document which services team members need to authenticate with
Consider creating a setup guide specific to your agent
This prevents confusion when team members first try to use the agent.
Combine with Workflow Triggers
For maximum automation, combine agents with Slack triggers:Agent: Handles ad-hoc questions in #support
Workflow Trigger: Automatically processes every new message for loggingThis gives you both interactive AI assistance and automated data capture.
For personal agents, make sure your personal default credentials are set. For workspace agents using workspace credentials, contact your workspace admin.
Wrong Agent Responding
Only one agent can be active per channel. If the wrong agent is responding:
Check active agent: /gummie active
Remove current agent: /gummie remove
Add correct agent: /gummie add [correct-agent-id]
Can't Add Agent to Channel
If you’re unable to add an agent:
✅ Verify channel type: Agents work in public and private channels, but not in DMs
✅ Check Gumloop is invited: Run /invite @Gumloop in the channel first
✅ Confirm permissions: Ensure you have permission to add apps to the channel
✅ Check agent ID: Verify you copied the complete agent ID from Gumloop
Agent Responses Too Slow
Agents may take longer when:
Calling multiple workflows
Using complex integrations
Processing large amounts of data
Solutions:
Optimize workflows to use fewer AI nodes
Limit agent tools to only what’s necessary
Use faster AI models in agent settings
Break complex tasks into smaller agent interactions
Can't Find Agent ID
To get your agent’s Slack command:Easiest method: Click “Connect to Slack” button in the agent menu and copy the commandAlternative:
Agent: Support Ticket AssistantChannel: #customer-support (public channel)Common interactions:
Copy
Ask AI
Team member: "@Gumloop Is ticket #12345 eligible for a refund?"Agent: [Checks ticket, reads policy, evaluates eligibility]Team member (in thread): "What about ticket #12346?"Agent (if "On all messages"): [Automatically checks ticket #12346]Or (if "Only on mentions"): [Waits for @mention]Team member: "@Gumloop Pull the last 5 interactions with [email protected]"Agent: [Searches CRM and email, returns history]
Why it works: Support team gets instant access to information without leaving Slack. Every interaction teaches the team how to use the agent for similar queries.Setup tip: Use “On all messages” mode for support channels to maintain conversation flow.
Agent: Lead Research AssistantChannel: #sales-team (public channel)Common interactions:
Copy
Ask AI
Team member: "@Gumloop Research this company: [LinkedIn URL]"Agent: [Enriches lead, checks CRM for activity, drafts outreach]Team member (in thread): "Thanks! What's the best time to reach out?"(With "Only on mentions" - no agent response, allows team discussion)Team member: "@Gumloop What's the status of opportunities over $50k?"Agent: [Queries Salesforce, summarizes pipeline]
Why it works: Sales reps get instant research without switching tools. Newer reps learn by watching experienced reps use the agent.Setup tip: Use “Only on mentions” mode to allow sales discussions without agent interruption.
Agent: Data Analysis AssistantChannel: #data-team (private channel)Common interactions:
Copy
Ask AI
Team member: "@Gumloop Compare Q4 revenue vs Q3 by product line"Agent: [Runs BigQuery workflow, generates comparison]Team member (in thread): "@Gumloop Now break it down by region"Agent: [Continues analysis with regional breakdown]Team member: "@Gumloop Flag any anomalies in yesterday's user signups"Agent: [Analyzes data, identifies outliers, reports findings]
Why it works: Analysts get quick answers to data questions. Non-technical team members learn what data is available and how to access it.Setup tip: Works in private channels too! Use “Only on mentions” for controlled analysis requests.