> ## Documentation Index
> Fetch the complete documentation index at: https://chatbotx.io/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Functions

> Learn how to configure AI Functions in ChatbotX so your Agent can look up data, process requests, and integrate with external applications automatically.

**AI Functions** let you connect AI with prebuilt automation workflows or external tools. Through AI Functions, your chatbot can collect data from users, trigger the right Flow to process a task, retrieve data from an outside API, and then return the result to AI so the conversation can continue naturally.

For example, if a user asks, *"What's the weather in Hanoi today?"*, AI alone will not have real-time weather data. But with an AI Function, you can trigger a Flow that calls a weather API, pass the result back to AI, and let it reply accurately.

## How to create an AI Function

To create a new one, go to **AI Functions** from the left navigation and click **+ Create AI Function**. The configuration screen includes these fields:

<Frame>
  <img src="https://mintcdn.com/chatbotx/6419eoAF_uo2OCBi/images/image-164.png?fit=max&auto=format&n=6419eoAF_uo2OCBi&q=85&s=67be59a7d686731e7952dbfd69fad469" alt="create_ai_function_button" width="3003" height="1181" data-path="images/image-164.png" />
</Frame>

* **Name**: The name of the AI Function. It should clearly describe what it does. For example, instead of a vague name like *"weather"*, use something like *"get\_current\_weather"*. It is best to use English or unaccented text, with at least 2 words.
* **Purpose (What does this trigger do?)**: This helps AI understand when the Function should be used. Explain its job clearly and in enough detail, such as *"Allows users to book a medical appointment"*. You do not need to list the required data in this field.
* **What data to collect?**: Configure the parameters AI needs to collect from the customer's replies before running the Flow, such as date, name, or email. **Note:** variable names should be meaningful. For example, use *"email"* instead of *"collect\_1"* so AI can map chat data correctly.
* **What is the output message?**: This is the raw result returned after the Function finishes running, such as *"Your appointment has been booked successfully"*. In most cases this message is not sent word for word. AI reads it, rewrites it into a smoother response, and sends that to the user.
* **Trigger Flow ID**: (*Optional*) Define the ID of the Flow you want to run after AI has collected all required data. Inside that Flow, you can save customer data, fire a webhook, call an API, or even hand the conversation over to a human with an Inbox Action.

<Frame>
  <img src="https://mintcdn.com/chatbotx/6419eoAF_uo2OCBi/images/image-167.png?fit=max&auto=format&n=6419eoAF_uo2OCBi&q=85&s=9ce25423830030174b57c71cfcb13126" alt="ai_function_configuration_modal" width="2101" height="1374" data-path="images/image-167.png" />
</Frame>

## Common AI Function examples

Below are some common examples to help you picture how an AI Function can be configured:

### 1. Book Appointment

* **Name**: `book_appointment`
* **Purpose**: Lets users book an appointment.
* **What data to collect?**: `date`, `email`, `reason`
* **What is the output message?**: Your appointment has been booked successfully.

### 2. Human Transfer

* **Name**: `connect_user_to_human`
* **Purpose**: Lets users talk to a real support staff member. This Function connects them directly.
* **Trigger Flow ID**: Triggers a Flow that contains an Inbox Action to hand the conversation to a human.
* **What is the output message?**: You have been connected to a support representative. We will reply shortly.

### 3. Weather API

* **Name**: `get_current_weather`
* **Purpose**: Lets users check the current weather in a specific location.
* **What data to collect?**: `city`
* **Trigger Flow ID**: Triggers a Flow that contains an External Request to call a weather API.
* **What is the output message?**: `{{data_from_api}}`

### 4. Email List

* **Name**: `join_email_list`
* **Purpose**: Lets users subscribe to our email newsletter list.
* **What data to collect?**: `email`
* **What is the output message?**: Subscription successful.

### 5. Shipping Quote

* **Name**: `request_shipping_quote`
* **Purpose**: Lets users request a shipping quote for a parcel.
* **What data to collect?**: `origin_address`, `destination_address`, `dimensions`, `weight`
* **Trigger Flow ID**: Triggers a Flow that automatically calculates the shipping quote based on parcel weight and size.
* **What is the output message?**: `{{quote}}`

## Using AI Functions

After setting up an AI Function, make sure it is enabled and linked to the correct AI Agent in that Agent's configuration.

<Frame>
  <img src="https://mintcdn.com/chatbotx/6419eoAF_uo2OCBi/images/image-158.png?fit=max&auto=format&n=6419eoAF_uo2OCBi&q=85&s=275d45281ebbc7b464e054a493feceb8" alt="ai_agent_select_knowledge_file" width="2098" height="1798" data-path="images/image-158.png" />
</Frame>

**Tip:** Your System Prompt can strongly influence how AI Functions behave. For example, if you have a restaurant booking Function, add a note like this to the Agent prompt: *"When a user wants to book a table, always ask for the expected time and the number of guests first."* That helps AI proactively collect all required data.