AI Agents are specialized assistants you can create and customize for specific tasks. While a regular chat draws on your entire knowledge base, an agent can be focused on a particular area, given specific instructions, and equipped with tools to get things done.
Agents listing page

​
What Makes an Agent Different from Chat?

Regular ChatAI Agent
Knowledge scopeSearches your full knowledge baseCan be scoped to specific documents or topics
InstructionsGeneral-purpose assistantCustom instructions for a specific role
ToolsStandard chat capabilitiesCan use specialized tools (search, connectors, etc.)
BehaviorOpen-ended conversationFocused on a defined task or workflow

​
Example Agents

Here are some agents your team might create:

HR Policy Assistant

An agent trained on your HR documents that answers questions about leave policies, benefits, and onboarding.

Sales Data Analyst

An agent with access to your CRM data and sales reports that can pull numbers and generate summaries.

Customer Support Helper

An agent scoped to your support documentation that drafts responses to common customer questions.

Research Assistant

An agent that combines your knowledge base with web search to compile research on any topic.

​
How Agents Work

1

Create the agent

Give it a name, description, and specific instructions about how it should behave.
2

Configure its capabilities

Choose which tools it can use, which knowledge it has access to, and how it should respond.
3

Start a conversation

Chat with the agent just like a regular conversation — but now it operates within the boundaries you defined.

​
Getting Started

Create an agent

Step-by-step guide to creating your first agent.

Workflow Builder

Use the visual builder for more advanced configurations.

Start from a template

Kickstart an agent from a ready-made template.

Manage agents

Edit, duplicate, and organize your agents.

Scheduled jobs

Run an agent automatically on a schedule.

API access

Trigger an agent from external systems over HTTP.