Accessing Agent Studio
To open Agent Studio:
Navigate to AI Agents from the left sidebar.
Click the Agent Studio tab.
The Agent Studio dashboard displays all agents in your account.
From this page, you can:
View existing agents
Check Draft or Published status
Organize agents into folders
Search for agents
Create a new agent
Access templates
Creating a New Agent
To create a new agent:
Click Create Agent.
Choose one of the following options:
Create from scratch
Browse templates
Use Create from scratch when building a completely custom agent. Use Browse templates when starting from a prebuilt structure.
Using the Template Library
The Template Library contains prebuilt agent blueprints designed for common use cases. Templates help you:
Launch faster
Learn how flows are structured
Reuse proven agent configurations
Modify existing logic instead of building from scratch
You can also:
Filter templates
Browse by category
Search for templates
Switch between grid and list views
Understanding the Builder Canvas
After creating or opening an agent, the visual builder canvas appears. The canvas is the main workspace where agents are designed. On the canvas, you can:
Drag and connect nodes
Configure execution logic
Add AI functionality
Create branching paths
Build structured conversations
Define how the agent starts and ends
Agents are built from left to right using connected nodes.
Adding a Start Trigger
Every agent requires a Start Trigger. The Start Trigger determines when the agent executes. Common trigger types include:
Form submitted
Chat message
Lead tag added
Without a Start Trigger, the agent cannot run. To add a Start Trigger:
Open the node panel.
Drag Start trigger onto the canvas.
Select the trigger type.
Understanding Node Categories
Nodes are the building blocks used to create agent behavior. Each node performs one specific task inside the flow. Core node categories include:
AI Agent
Sequential
Start trigger
End Node
Router
Search Knowledge Base
Search the web
MCP Server
Use these nodes to control execution flow, retrieve information, and connect tools.
AI and Capture Nodes
Agent Studio also includes nodes for AI generation and structured data collection. Available nodes include:
AI & Content Nodes
API Call
Audio Generation
Image Generation
Text Generation
Video Generation
Information Capture Nodes
Email Address
Phone Number
Single Choice
Text Input
These nodes allow agents to generate content and collect user responses during interactions.
Variables
Variables store reusable information across the agent. Use Variables when you need the same value referenced in multiple places. Common examples include:
Business name
Support email
Product name
Dynamic user information
Variables help maintain consistency throughout the flow.
Global Prompt
The Global Prompt defines instructions that apply across the entire agent. Use the Global Prompt to control:
Tone of voice
Agent personality
Response rules
Formatting behavior
AI restrictions or boundaries
The Global Prompt influences how the AI behaves during every interaction.
Testing an Agent
Before publishing, use the Test feature to simulate the agent experience. Testing helps confirm:
Triggers work correctly
Nodes execute in the expected order
AI responses follow instructions
User inputs are captured properly
Routing logic behaves correctly
Testing should always be completed before publishing.
Publishing an Agent
Agents must be published before they can respond to real-world events.
To publish an agent:
Save the flow.
Click Publish.
Confirm the publication.
Draft agents do not execute outside of testing mode.
How Agent Studio Works
A typical agent flow follows this structure:
A trigger starts the agent
The agent processes information
The flow collects or generates data
Logic determines the next action
The flow reaches an End Node
This structure allows agents to behave dynamically instead of following rigid automation rules.
How to Create Your First Agent
Navigate to AI Agents → Agent Studio
Click Create Agent
Choose Create from scratch or Browse templates
Add a Start Trigger
Add an AI Agent node
Connect the nodes together
Add additional logic or tools
Use Test to simulate the flow
Save the agent
Click Publish
Common Issues and Fixes
The agent is not running
Confirm the following:
A Start Trigger is configured
The trigger conditions are valid
The agent has been published
Draft agents do not respond to real events.
The agent is responding incorrectly
Review the following:
Global Prompt instructions
AI node configuration
Routing conditions
Variable values
More specific instructions generally improve response accuracy.
The flow stops unexpectedly
Check for:
Unconnected nodes
Missing routes
Missing End Nodes
Invalid logic conditions
All execution paths should lead to a valid completion point.
The agent is not retrieving the correct information
Verify:
Knowledge Base connections
Search node configuration
API responses
Variable mappings
Incorrect configuration can prevent the agent from accessing expected data.
FAQ
Do agents run automatically?
Agents run only when their configured trigger conditions are met.
How is Agent Studio different from Workflows?
Workflows follow predefined automation rules. Agent Studio uses AI to interpret input, make contextual decisions, and dynamically route conversations.
Does an agent need to be published before it works?
Yes. Agents must be published before they can respond to live events.
Can Agent Studio connect with external systems?
Yes. API Call and MCP Server nodes allow integrations with external systems and tools.
Can agents generate content other than text?
Yes. Agent Studio supports text, image, audio, and video generation.
Can I modify a template after installing it?
Yes. Templates are fully customizable after installation.