Forms

Forms (also called Data Types) let you define structured data for your workspace. AI agents can fill them through natural conversation, or team members can enter data manually.

Creating a Form

1. Go to Knowledge → Forms
2. Click Create Form
3. Fill in the form settings:

Field Types

Each form contains fields that define what data to collect. Add fields by clicking Add Field in the form editor.

Field Settings

Every field has these configurable settings:

Dropdown and Multi-select Options

For Dropdown and Multi-select fields, you define a list of options:

• Label — What's displayed to the user
• Value — What's stored (can differ from the label)
• Visible When — Optional CEL expression to conditionally show/hide an option based on other field values (e.g., record.data.category == "urgent")

You can add options one at a time or use Bulk Mode to paste multiple options at once.

Collections vs. Single Records

The Collection toggle changes how the form behaves:

In the UI, single-record forms show as a form view while collections show as a table of entries.

Using Forms in Workflows

Attach forms to tasks in your workflows so your AI agent can collect data naturally through conversation.

1. Create or edit a Task in the workflow editor
2. In the task settings, attach a form under Data Collection
3. The AI uses the form's Prompt field for context on how to collect the data
4. During conversation, the AI asks for each field and validates responses automatically

The AI adapts to the conversation flow — it doesn't rigidly go field by field. It can collect multiple values from a single message and ask follow-up questions when needed.

Validation

Beyond basic required/type validation, you can add custom validation rules using CEL expressions. Set a validateCel expression on any field to enforce business rules.

Available variables in validation expressions:

The expression should return true if valid, or a string error message if invalid.

cel

Bulk Import & Export

Import records in bulk from CSV files:

1. Open the form and go to the Records tab
2. Click Import and upload a CSV file
3. Map CSV columns to form fields
4. Review and confirm the import

The bulk importer supports up to 10,000 records per batch. It can create new records or update existing ones using an external ID for matching.

To export, click Export on the Records tab to download all records as CSV.

Events & Automations

Forms trigger events you can use with Actions:

Use these events to trigger notifications, webhooks, or other automated actions. Target specific forms using conditions in your action rules.

Referencing Form Data

Access form data in template variables and messages:

{{member.data.<form_slug>.<field_slug>}}

For single-record forms, this returns the field value directly.

For collection forms, use member.collections for structured access:

{{member.collections.<form_slug>.count}}
{{member.collections.<form_slug>.latest.data.<field_slug>}}
{{member.collections.<form_slug>.latest.created_at}}

Loop through recent records (up to 10):

{% for record in member.collections.<form_slug>.recent %}
  - {{record.data.<field_slug>}} ({{record.created_at}})
{% endfor %}

See the Template Variables guide for more details.

Tips

• Use clear prompts — The form's Prompt field guides the AI. Be specific about tone, order of questions, and validation expectations.
• Test with conversation — After setting up a form, test it by chatting with your agent to make sure the data collection flow feels natural.
• Index searchable fields — Enable search indexing on fields you'll frequently filter or search by. This runs a background indexing job.
• Slugs are permanent — Plan your field slugs carefully. Changing a slug after data has been collected can break references and lose data.

Related

• Knowledge — Overview of all Knowledge features
• Computed Fields — Auto-calculated formula fields
• Actions — Trigger automations from form events
• Template Variables — Use form data in messages