> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getinteract.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Building agents

> Create and configure an AI agent tailored to your data, your questions, and your team's way of working.

## Before you start

Make sure you have at least one [BigQuery connector](/org-settings/connectors/google-big-query) set up in **Settings → Connectors**. Your agent needs a connector to query live data.

## Creating a new agent

<Steps>
  <Step title="Go to Agents">
    From the left sidebar, click **Agents** and then **Create agent**.
  </Step>

  <Step title="Set a name and description">
    Give your agent a clear, descriptive name — for example, *"Paid Search Analyst"* or *"Campaign Budget Tracker"*.
  </Step>

  <Step title="Choose an agent type">
    Select **Data Analysis Agent** for most use cases. Use **Multi-Agent Swarm** if you need separate, independently configured subagents. See [General settings](/agents/settings/general#agent-type) for a comparison.
  </Step>

  <Step title="Configure the Agent Builder">
    Write the orchestrator system prompt — tell the agent what it does, which data it works with, and how to format its responses. [Agent Builder →](/agents/settings/agent-builder)
  </Step>

  <Step title="Set up data access">
    Connect a data source, restrict which datasets and tables the agent can query, and control which users can access it. [Data Access →](/agents/settings/data-access)
  </Step>

  <Step title="Save and test">
    Click **Save**, then **Open chat** to test your agent. Adjust the system prompt and context until the results match your expectations.
  </Step>
</Steps>

## Agent settings

Each agent has five configuration tabs. Click any card below to jump to the detailed reference.

<CardGroup cols={2}>
  <Card title="General" icon="sliders" href="/agents/settings/general">
    Name, slug, agent type, model, and display settings.
  </Card>

  <Card title="Agent Builder" icon="diagram-project" href="/agents/settings/agent-builder">
    Orchestrator prompt, subagents, tools, model overrides, and context.
  </Card>

  <Card title="Data Access" icon="database" href="/agents/settings/data-access">
    Data source, connector, dataset and table filters, and user group restrictions.
  </Card>

  <Card title="Guardrails" icon="shield-check" href="/agents/settings/guardrails">
    Input and output checks: PII masking, off-topic detection, custom rules.
  </Card>

  <Card title="Instructions & UX" icon="pen" href="/agents/settings/instructions">
    Agent logo, start-page instructions, and example prompts.
  </Card>
</CardGroup>

## Testing your agent

Once saved, click **Open chat** to test your agent. Check:

* Does it query the right tables?
* Does it use the correct date ranges and filters?
* Is the output format what you expected?

Each time you save, a new prompt version is created — you can always roll back to a previous version.

[Learn about prompt versioning →](/agents/prompt-management)