What is AI Form for?

The AI Form node is used when your AI agent needs to collect several pieces of structured information from a user through a guided conversation.

Questions are asked one by one in order. Each answer is processed and saved before the next question is shown.

AI Form is useful for:

• Lead intake
• Customer qualification
• Onboarding
• Applications
• Booking or service requests
• Support intake
• Quote requests
• Any workflow where several answers need to be collected reliably

What AI Form does

AI Form lets you configure multiple questions inside one node.

For each question, you can control how the answer is collected, checked, cleaned, and saved.

AI Form can:

• Ask multiple questions one at a time
• Save each answer before moving to the next question
• Require a question so the user cannot skip it
• Add quick-reply buttons for common answers
• Validate or verify answers before accepting them
• Use AI Verification for custom answer criteria
• Use AI Answer Cleanup to extract a clean value before saving
• Give each saved answer a descriptive label
• Show a confirmation summary before submission
• Let the user correct only the field that is wrong
• Send a completion message after the form is finished
• Email collected responses to selected recipients
• Make all collected answers available as context for later AI nodes

How AI Form works

AI Form follows a clear sequence:

1. The AI asks the first question.
2. The user replies.
3. If verification is configured, the answer is checked.
4. If the answer does not pass, the AI asks again.
5. If verification passes, optional answer cleanup runs.
6. The processed value is saved.
7. The AI moves to the next question.
8. This continues until all questions are completed.
9. If confirmation is enabled, the user reviews a summary.
10. The collected answers become available to nodes that come after the form.

For questions that use both AI Verification and AI Answer Cleanup, the execution order is:

User answers → AI Verify → AI Cleanup → Saved

Add questions

Inside AI Form Settings, click + Add Question to add the questions you want the AI agent to ask.

Each question has its own settings.

For example, a lead form could ask:

1. What is your full name?
2. What is your work email?
3. What company do you work for?
4. What are you interested in?
5. What is your budget?

The AI asks them in the order they appear.

Can't skip

Enable Can't skip when the user must provide an answer before the form can continue.

Use this for information that is required for the workflow.

For example:

• Email address
• Policy number
• Appointment preference
• Required qualification information

Leave it off when the question is optional.

Quick Reply Buttons

Each question can include optional Quick Reply Buttons.

These let the user tap a predefined answer instead of typing.

You can add up to 3 buttons, with up to 20 characters per button.

Quick replies work well for questions such as:

• Yes / No
• Product category
• Service type
• Preferred location
• Customer type

They can make forms faster and reduce ambiguous answers.

Validation and AI Verification

Validation is handled directly inside AI Form. You do not need a separate Validation node.

For each question, choose the appropriate Validation Rule.

If no validation is needed, leave the rule as None.

AI Verification

Choose AI Verification when the answer needs to be checked against custom criteria.

You provide a Verification Prompt that describes what a valid answer should look like.

For example:

1Check that the answer is a valid business email and not a personal address such as Gmail or Yahoo.

The AI checks the user's answer against that instruction.

If the answer does not pass, the question is asked again.

You can also enable Show message when validation fails and provide a custom message such as:

1That doesn't look right. Please try again.

AI Verification is useful when correctness depends on meaning or business rules rather than only a fixed format.

Answer Label

Each question has an Answer Label.

The label names the field in the collected data.

Use descriptive labels such as:

• Full Name
• Work Email
• Company
• Budget
• Product Interest
• Destination

A clear Answer Label is especially important when using AI Answer Cleanup because it helps the AI understand what value it should extract.

AI Answer Cleanup

Enable AI Answer Cleanup when you want AskHandle to quietly turn a conversational response into a clean stored value.

For example:

1User reply:
2My name is John Smith
3
4Saved value:
5John Smith

Cleanup happens after verification passes and does not re-prompt the user.

Normalize text

Use Normalize text when you want to extract or clean up the user's answer while preserving the intended value.

This works well for:

• Names
• Emails
• Phone numbers
• Locations
• Product names
• Dates
• Other free-text values

Map to fixed values

Use Map to fixed values when the saved answer should be normalized into one of a defined set of values.

This is useful when users may describe the same answer in different ways but your workflow needs a consistent stored value.

Extraction Hint

When AI Answer Cleanup is enabled, you can optionally add an Extraction Hint.

Use it when the question could be ambiguous or when you want the AI to extract a very specific part of the answer.

Example:

1Extract only the color word, such as red, blue, or white.

For common fields such as name, email, or phone, an extraction hint may not be necessary.

Confirmation step

Under After questions are answered, you can enable:

Ask user to confirm before submitting

When enabled, AI Form shows the user a summary of the collected answers.

The user can review the information before the form is completed.

If the user says one field is wrong, the AI corrects only that field.

For example:

Actually, my email should be alex@newcompany.com.

The AI updates the email answer without restarting the entire form.

This is useful for forms where accuracy matters, such as:

• Lead qualification
• Applications
• Booking requests
• Quote requests
• Service intake

Completion message

You can add an optional Completion message that is sent immediately after all questions are answered.

Example:

1You're all set! Is there anything else I can help you with?

If you leave the completion message empty, the form can finish silently and continue to the next node.

Email me responses

AI Form can optionally send the collected answers by email every time the form is completed.

Under Email me responses, enter one or more recipient email addresses.

This can be useful when a team wants immediate visibility into:

• New leads
• Applications
• Service requests
• Booking inquiries
• Intake forms

Multiple recipient addresses can be added as supported in the field.

Using collected answers later in the flow

All answers collected by AI Form are available as context for AI nodes that come after it.

For example:

1AI Form → AI Answer

AI Form might collect:

• Name
• Company
• Budget
• Requested service

The next AI Answer node can then use those values to generate a response that is specific to that user.

This means you do not need to ask the same questions again later in the flow.

AI Form vs. Question node

Use AI Form when you need to collect multiple pieces of information in one structured intake.

Use the Question node when you only need to ask one question and wait for the user's reply.

The Question node can also validate the reply when a specific format is required, such as:

• Email
• Phone number
• URL
• List / CSV lookup
• Custom pattern

A simple rule:

Several structured questions → AI Form

One validated question → Question node

AI Form vs. a normal AI conversation

A normal AI conversation is better for open-ended discussion.

AI Form is better when you need reliable, separate fields that can be reused later.

Use AI Form when:

• Specific information must be collected
• Answers need to be saved separately
• Some questions are required
• Answers need verification or cleanup
• Users should review their answers before submission
• The collected data will be used by later nodes or business processes

Example: lead qualification form

An AI Form could collect:

1. Full name
2. Work email
3. Company
4. Number of employees
5. What the customer needs help with
6. Estimated budget

You could configure:

• Can't skip for name and email
• AI Verification to check that the email meets your criteria
• AI Answer Cleanup to store clean values
• Quick replies for company size
• Confirmation before submission
• Email delivery to your sales team

After the form finishes, an AI Answer node can use those collected answers to continue the conversation.

Best practices

Ask only what you need

Long forms create more friction. Collect only the information that is necessary for the next step.

Use clear Answer Labels

Good labels make saved data easier to understand and improve AI Answer Cleanup.

Use AI Verification selectively

Add verification when a wrong or incomplete answer would affect the workflow.

Use cleanup for natural replies

AI Answer Cleanup is useful when users are likely to answer conversationally rather than with a clean value.

Use quick replies when choices are predictable

Buttons make simple questions faster to answer and reduce variation.

Use confirmation for important submissions

Confirmation is valuable when incorrect information could affect a booking, application, lead, quote, or service request.

Test the complete form

Test valid answers, invalid answers, corrections, skipped questions, cleanup behavior, and confirmation before publishing.

Important to know

• Question Flow no longer exists.
• Use AI Form for multi-question structured intake.
• Validation and AI Verification are configured directly inside AI Form.
• AI Answer Cleanup runs after verification and before the value is saved.
• Answers are saved one at a time before the next question is shown.
• The confirmation step lets users correct individual fields without starting over.
• Collected answers remain available as context for later AI nodes.
• Use the Question node when you only need one validated question.