Skip to content

Add a form agent

4 min read

Create a form agent that guides a user through filling out a form, collecting structured answers that match a schema you define.


A form agent helps a user fill out a form through conversation. Instead of showing static fields, it asks questions one at a time and collects structured answers matching a schema you define — combining a chat experience with the structured output of an extraction agent.

Agents live in your workspace and are created from the New Agent card on the workspace overview (or the + next to Agents in the left sidebar). Choosing the Form type is what turns the agent into a guided form.

How a form agent works

You define the fields to collect on the Form tab (for example full_name and email), a Greeting the agent opens with, and Instructions describing how it should guide the user. A form agent has three configuration tabs — General, Model and Form — and each tab is its own form that saves on its own (the editor warns you before you leave a tab with unsaved changes).

The full flow at a glance

The walkthrough below replays every step in the real interface. Use Prev / Next to move at your own pace; each step highlights the button to click and the area to watch.

DemoStudio
Conversation Bayes Assistant
Form Form Agent
New Agent

An agent can perform specific tasks based on its type and configuration.

Create
Edit Form AgentEdit details for this form agent
General Model Form

Name

Contact Form

Locale

EN

Greeting

Hi! I'll help you fill out this form. Let's get started.

Instructions

Your main task is to help the user fill out the form by asking questions and providing guidance. Ask one question at a time.
Save
Edit Form AgentEdit details for this form agent
General Model Form

Model

Select model

Temperature

0.2
Save
Edit Form AgentEdit details for this form agent
General Model Form
Advanced mode
Ask questions in this order
#Field nameTypeRequired
1 full_name Text
2 email Text
Add field
Save

Step by step

1. Open the New Agent dialog

On the workspace overview, find the New Agent card and click Create (you can also use the + next to Agents in the left sidebar).

2. Choose the Form type

In the New Agent dialog, under Agent type, select Form. The type determines what the agent does, so pick it before continuing.

3. Name the agent and create it

Type a Name (at least 3 characters) in the Name field, then click Create. Your agent is created with sensible defaults and opens in the editor.

4. Set the basics on General

The editor opens on the General tab. Set the:

  • Name — how the agent appears in your workspace.
  • Locale — the language the agent works in.
  • Greeting — the first message the agent sends when a conversation starts.
  • Instructions — how the agent should guide the user through the form.

Click Save.

5. Pick the model on Model

Open the Model tab, choose the Model that powers the agent and a Temperature (lower values give steadier, more predictable questions). Save.

6. Build the form on Form

Open the Form tab and build the fields the agent collects visually — no JSON required. Click Add field and, for each field, set:

  • Field name — the key the answer is stored under (e.g. full_name).
  • TypeText, Number, Yes / No, List, or Choice. Number offers an optional min/max range, and Choice lets you define the allowed values.
  • Description — the question or hint the agent uses to ask for this field.
  • Required — toggle on for mandatory answers.

Remove a field with its trash icon.

To control the order in which the agent asks, turn on Ask questions in this order and drag the rows into the order you want; turn it off to let the assistant choose the order itself. Need finer control? Click Advanced mode (top right) to edit the raw JSON, and Visual editor to switch back. Click Save — and remember each tab saves on its own.

Version history

Every save creates a new version of the agent. Open History (top right of the editor, showing the current v number) to browse past versions in a side panel, compare what changed (Changes in this version or Compare with current), and Restore this version — which copies those settings as a new current version without deleting anything. See Version history for the full walkthrough.

Tips

  • Keep field descriptions clear — the agent uses them to phrase each question and to decide what to store.
  • Use Ask questions in this order when the sequence matters (e.g. ask for a name before a preference); leave it off to let the agent adapt.
  • The Greeting sets the tone; write it in the same language as the agent’s Locale.
  • Use a low Temperature (for example 0.2) so the agent asks consistent questions rather than improvising.
  • Each tab saves separately — changes on one tab aren’t kept when you move to another without saving.

Troubleshooting

  • The Create button is disabled — the Name must be at least 3 characters.
  • I don’t see the field type I need — the visual editor covers Text, Number, Yes / No, List and Choice; use Advanced mode (raw JSON) for nested objects.
  • The agent asks questions in the wrong order — turn on Ask questions in this order and drag the fields into the sequence you want.
  • My changes disappeared — each tab saves separately; click Save before switching tabs (the editor warns you when changes are unsaved).
  • I don’t see a Form tab — make sure you created a Form agent (conversation agents chat, extraction agents show an Output tab instead).

Last updated: July 22, 2026

Was this article helpful?