Platform Actions

Actions represent the core set of operations that your conversational AI agents can execute during interactions. Unlike abilities which are configurable capabilities, actions are predefined operations that handle common conversational tasks such as responding to users, asking questions, retrieving information, or triggering workflows.

Understanding Platform Actions

The platform provides a curated set of actions that have been designed to work seamlessly with the conversational AI framework. Each action is carefully defined with clear semantics and expected behaviors, ensuring consistent and reliable operation across different contexts and use cases.

To retrieve the complete catalog of available platform actions:

Each action in the response includes:

• id: Unique identifier for the action
• name: Standardized name following action naming conventions
• description: Clear explanation of what the action does and when to use it
• examples: Real-world examples demonstrating how to invoke the action

Action Examples and Usage Patterns

The examples array is particularly valuable as it provides concrete demonstrations of how each action should be used in practice. These examples often reference specific abilities (indicated by the @ prefix) and show the exact syntax and structure needed to invoke the action correctly.

Building with Actions

When designing conversational flows or building agent behaviors, use actions as the building blocks for your logic. Actions provide a structured approach to defining what your agent should do at each step of a conversation. By combining multiple actions in sequence or conditionally, you can create sophisticated conversational experiences.

Review the examples provided with each action to understand typical usage patterns and best practices. The examples often demonstrate how actions integrate with abilities to create complete functionality, such as combining a search action with a web search ability to retrieve and present information.

Important: Actions are designed to work within the platform's conversational framework and may have dependencies on specific abilities or platform features. Always test your action implementations thoroughly to ensure they behave as expected in your specific use case.