Which AskHandle node should I use?

AskHandle includes different nodes for different jobs inside an AI agent.

The right node depends on what you want the AI agent to do.

For example:

• Use AI Answer for conversational answers, instructions, and AI skills
• Use Document Search to answer from documents and written knowledge
• Use Data Search to search structured data such as CSV or Excel files
• Use AI Form to collect several pieces of structured information
• Use Question when you need to ask one specific question
• Use Message when you need to send a fixed message
• Use Router when different customer requests should go to different parts of the workflow
• Use Human Handoff when the conversation needs to move to a person

You can combine these nodes inside the same AI agent.

Quick guide

AI Answer

Use AI Answer when you want the AI agent to understand a customer's request and respond conversationally.

AI Answer can use the instructions and knowledge you configure and can also have additional skills enabled.

Depending on your setup, AI Answer can be used for tasks such as:

• Answering general customer questions
• Explaining products or services
• Providing guidance
• Using business-specific instructions
• Searching live websites
• Performing calculations
• Checking availability and booking appointments
• Sending email
• Showing products or images

A simple AI agent may only need:

Start → AI Answer

As your use case becomes more complex, you can add other nodes around it.

Example

A customer asks:

“Can you help me choose the right service?”

If the answer depends mainly on conversation, instructions, and general business knowledge, AI Answer is usually the right choice.

Document Search

Use Document Search when the answer should come from written knowledge stored in documents.

Common sources include:

• Policies
• Manuals
• FAQs
• Guides
• Help articles
• Service information
• Internal documentation
• PDF documents

Document Search is useful when the information is mostly written as paragraphs, pages, or articles rather than rows and columns.

Example

A customer asks:

“What is your cancellation policy?”

If the policy is stored in an uploaded document or knowledge article, Document Search can find the relevant information and answer from it.

Simple rule

If the information reads like a document, use Document Search.

If the information looks like a spreadsheet, use Data Search.

Data Search

Use Data Search when the AI agent needs to search structured information.

Data Search works with information organized into rows and columns, such as CSV or Excel files.

Common examples include:

• Product catalogs
• Prices
• Inventory
• Customer records
• Policies
• Orders
• Locations
• Insurance data
• Reference tables

Example

A customer asks:

“How much does product X cost?”

If your product information is stored in a spreadsheet containing product names, SKUs, prices, and availability, use Data Search.

Data Search can work with a single file or with multiple related datasets.

AI Form

Use AI Form when you need to collect several pieces of structured information from a user.

AI Form asks questions one by one and saves the answers into separate fields.

It is useful for:

• Lead qualification
• Quote requests
• Service requests
• Applications
• Customer intake
• Booking information
• Support intake
• Collecting contact information

AI Form can also use AI Verification to check an answer before accepting it and AI Answer Cleanup to normalize or extract the value that should be saved.

Example

You need to collect:

• Name
• Email
• Service needed
• Location
• Preferred date

Instead of asking everything in one message, AI Form can collect each answer step by step.

Question

Use Question when you only need one specific piece of information.

For example:

• “What is your email address?”
• “What is your order number?”
• “Which location are you contacting?”
• “What is your vehicle's engine power in kW?”

The answer can be saved and used later in the workflow.

If you need to collect several fields, use AI Form instead.

Simple rule

One question → Question

Several structured questions → AI Form

Message

Use Message when the AI agent needs to send a specific message without generating a new AI response.

This is useful for:

• Instructions
• Confirmations
• Notices
• Introductions
• Transition messages
• Fixed information

Example

After a customer completes an intake process, a Message node could say:

“Thank you. We have received your information and will connect you with our team.”

Use Message when the wording should remain controlled rather than generated dynamically.

Router

Use Router when your AI agent has several possible paths and needs to decide where each customer message should go.

Router identifies what the user is trying to do and sends the conversation to the appropriate connected node.

For example:

Pricing question
→ Data Search

Policy question
→ Document Search

Book an appointment
→ AI Answer with Scheduling enabled

Speak to someone
→ Human Handoff

Router is useful when one AI agent handles several different types of requests.

You usually do not need Router when your AI agent only has one main path.

Simple rule

One job or path → Router may not be necessary

Several different jobs → Use Router to direct the conversation

Human Handoff

Use Human Handoff when the conversation needs to move from the AI agent to a person.

This can be useful when:

• The customer asks to speak with someone
• The request requires human judgment
• Your team needs to review the case
• The AI cannot complete the request
• A business process requires human involvement

Human Handoff can be used after other nodes.

For example:

Router → AI Form → Human Handoff

The AI agent can first understand the request and collect the required information before transferring the conversation.

Nodes can work together

You do not have to choose only one node for an AI agent.

Most advanced agents combine nodes.

For example, a customer service AI agent could use:

1. Router to understand the request
2. Document Search for policy questions
3. Data Search for account or product information
4. AI Form to collect required details
5. AI Answer for conversational assistance or skills
6. Human Handoff when a person needs to take over

Each node should have a clear job.

Example: appointment request

A customer says:

“I'd like to book an appointment.”

A workflow might use:

Router
Identifies that the customer wants to book.

↓

AI Answer with Scheduling enabled
Checks availability and helps the customer choose a time.

The customer does not need to know that different nodes or skills are working behind the conversation.

Example: support request

A customer says:

“My order arrived damaged.”

A workflow could use:

Router
Recognizes a support request.

↓

AI Form
Collects the order number, email address, and problem details.

↓

Human Handoff
Transfers the conversation to the support team with the collected information.

Example: product question

A customer asks:

“Do you have this model available in black?”

The AI agent could use Data Search if product availability is stored in a spreadsheet or structured product dataset.

If the customer then asks:

“Which model would you recommend for a small office?”

The workflow may use AI Answer for the conversational recommendation.

Different nodes can handle different parts of the same customer journey.

How to choose the right node

Start by asking what the AI agent needs to accomplish.

The customer needs an answer

Use:

• AI Answer for conversational responses and AI skills
• Document Search for written knowledge
• Data Search for structured records or values

You need information from the customer

Use:

• Question for one answer
• AI Form for several structured answers

You need to control the workflow

Use:

• Message to send fixed information
• Router to choose between different paths

A person needs to continue the conversation

Use:

• Human Handoff

Keep workflows as simple as possible

More nodes do not automatically make an AI agent better.

If one AI Answer node can perform the job reliably, you may not need a complex workflow.

Add additional nodes when they solve a specific need, such as:

• Searching a dedicated knowledge source
• Looking up structured data
• Collecting information
• Separating different customer intents
• Transferring the conversation to your team

Clear responsibilities make AI agents easier to build, test, and maintain.

Related guides

For more detailed setup instructions, see the individual Help Center guides for:

• AI Answer
• Router
• Document Search
• Data Search
• AI Form
• Message and Question
• Human Handoff
• AI Answer skills