Track Button Method

The track_button() method allows you to log user interface interactions, specifically button clicks and selections. This provides insights into how users interact with your AI agent's interface elements.

Method Signature

def track_button(
    button_event: str,
    label: str,
    value: str,
    conversation_id: Optional[str] = None,
    metadata: Optional[Dict[str, Any]] = None
) -> Dict[str, Any]:

Parameters

Parameter	Type	Required	Description
button_event	str	Yes	Description of what this button is used for
label	str	Yes	The button label/text displayed to the user
value	str	Yes	The button value
metadata	Dict[str, Any]	No	Additional contextual information

Complete Usage Example

E-commerce Button Tracking

# Track product selection button
tracker.track_button(
    button_event="product_selection",
    label="Add to Cart",
    value="product_123",
    metadata={
        "product_name": "Wireless Headphones",
        "price": 99.99,
        "category": "electronics"
    }
)

Support Ticket Actions

# Track support resolution button
tracker.track_button(
    button_event="ticket_resolution",
    label="Mark as Resolved",
    value="resolved_satisfied",
    metadata={
        "resolution_type": "self_service",
        "satisfaction_rating": 5,
        "issue_category": "billing"
    }
)

Navigation Tracking

# Track navigation button clicks
tracker.track_button(
    button_event="navigation_action",
    label="View Order History",
    value="order_history_page",
    metadata={
        "section": "account_management",
        "user_type": "premium_member"
    }
)

Track which buttons were clicked

You may want to render multiple buttons which were rendered but only one was clicked. In order to do that you need to pass the same button_event but different label and value. AgentSight automatically tracks different buttons with the same button_event and renders them inside the conversation preview in the Dashboard.

Example

Track your buttons when user asks for contacting human agent:

tracker.track_button(
    button_event="contact_human_agent",
    label="Copy conversation",
    value="copy-conversation"
)

tracker.track_button(
    button_event="contact_human_agent",
    label="Create ticket",
    value="create-ticket"
)

Rendering inside conversation preview

Only track clicked buttons, in the conversation preview, AgentSight will automatically pull all the buttons with the same button_event.

Use Cases

• User Experience Analytics: Track which interface elements users interact with most frequently to optimize your UI design.
• Conversion Tracking: Monitor button clicks that lead to important actions like purchases, sign-ups, or support requests.
• A/B Testing: Compare the performance of different button labels or placements by tracking interaction rates.
• User Journey Mapping: Understand how users navigate through your AI agent's interface and identify common paths.

Important Notes

Button Context

Button interactions are tracked within the context of a conversation. Use the same conversation_id across related interactions to maintain proper context.

Required Fields

All three core fields (button_event, label, and value) are required and cannot be empty. The method will raise an exception if any of these fields are missing or empty.

Best Practices

• Use descriptive button_event names that clearly indicate the button's purpose
• Keep label text consistent with what users actually see in the interface
• Use meaningful value strings that help identify the specific action or selection
• Include relevant metadata to provide additional context for analytics