Why should I use Dispatcher?

Use Dispatcher when your AI agent needs to understand what a customer is asking and send them to the right next step.

Dispatcher is a real-time routing node. It monitors every user message, uses fast AI to detect the customer’s intent, and then routes the conversation through the correct output connection.

This is useful when your agent can handle more than one type of request, such as pricing questions, support questions, document questions, or lead collection.

What Dispatcher does

Dispatcher helps your AI agent decide where the conversation should go next.

Instead of sending every customer message to the same node, Dispatcher checks the message, identifies the intent, and routes it to the matching path.

Example:

• A customer asks, “How much does this cost?” → Dispatcher routes them to a pricing path.
• A customer asks, “What is your cancellation policy?” → Dispatcher routes them to a document or knowledge path.
• A customer says, “Hi, I need help” → Dispatcher routes them to the fallback path.

What is an intent?

An intent is what the customer is trying to ask or do.

Each intent you create becomes an output handle on the Dispatcher node. After saving, you can drag that handle and connect it to the next node in the workflow.

Example intents:

• Asking Price
• Book Appointment
• Product Question
• Refund Request
• Collect Contact Info
• General

How Dispatcher routes messages

Dispatcher continuously watches the conversation and checks each new customer message.

When a message matches one of your intents, Dispatcher sends the conversation through that intent’s output edge.

For example:

1Intent: Asking Price
2Trigger: User is asking about pricing, costs, or how much something costs.
3Connected to: Data Search

If the customer asks, “What are your prices?”, Dispatcher sends the conversation to the Data Search node connected to the Asking Price intent.

Using Dispatcher with Message nodes

Specific intents can also route to Message nodes.

Use this when you want to collect information from the customer before continuing.

Common examples:

• Collect email address
• Collect phone number
• Ask for company name
• Ask for preferred appointment time
• Ask for order number

Example:

1Intent: Book Appointment
2Connected to: Message node asking for phone number

This lets your agent guide the customer step by step instead of giving the same response to everyone.

How to set up Dispatcher

1. Connect Start to Dispatcher

Connect the Start node to the Dispatcher node.

This makes Dispatcher the first decision point in the workflow.

2. Add your intents

Open the Dispatcher settings and click Add Intent.

Each intent has two fields:

Name
A short label for the route.

Example:

1Asking Price

Intent Trigger
A clear description of what the customer may ask.

Example:

1User is asking about pricing, costs, or how much something costs.

The customer does not need to use the exact same words. Dispatcher uses AI to understand the meaning of the message.

3. Keep the fallback as the last intent

The last intent is always the fallback.

Fallback handles general conversation, unclear messages, greetings, or anything that does not match your specific intents.

Example:

1Name: General
2Intent Trigger: General conversation or unclear intent.

New intents are added above the fallback so the fallback stays last.

4. Save your changes

Click Save & Close after adding or editing your intents.

5. Connect each intent handle to the next step

After saving, each intent appears as a handle on the Dispatcher node.

Drag each intent handle to the node that should handle that request.

Example:

• Asking Price → Data Search
• Product Question → Document Search
• Book Appointment → Message node
• General → fallback support path

This connection tells Dispatcher exactly where to send the customer next.

Best practices

Keep each intent focused

Create one intent for one clear type of request.

Good:

1User is asking about pricing, costs, or how much something costs.

Avoid combining unrelated topics, such as pricing, booking, and refunds, into one intent.

Write intent triggers as full descriptions

Avoid one-word triggers.

Instead of:

1Pricing

Use:

1User is asking about pricing, costs, or how much something costs.

Clear descriptions help Dispatcher route messages more accurately.

Always use the fallback

Customers often ask questions in unexpected ways.

The fallback keeps the conversation moving when Dispatcher cannot confidently match a specific intent.

When to use Dispatcher

Use Dispatcher when your AI agent has multiple possible conversation paths.

Common examples:

• Routing pricing questions
• Routing document or policy questions
• Routing product questions
• Collecting lead details
• Booking appointments
• Sending support requests to the right flow
• Handling general conversation with a fallback

If your agent only performs one simple task, you may not need Dispatcher. If your agent needs to choose between different next steps, Dispatcher helps keep the workflow clear and reliable.

Key takeaway

Dispatcher is the decision-making layer of your AI agent.

It monitors each user message, detects the customer’s intent in real time, and sends the conversation to the correct next step.

Use it to build cleaner workflows, guide customers faster, and make sure each request is handled by the right node or flow.