Einstein Copilot is a conversational AI assistant that helps Salesforce users execute commands using natural language. When developers create
custom actions using Apex methods annotated with @InvocableMethod, Copilot relies on the label and description
attributes to understand what each method does.
Without clear and descriptive labels and descriptions, Einstein Copilot cannot effectively match user conversation intents with the appropriate
actions. This leads to poor user experience where:
- Users' natural language requests may not trigger the intended custom actions
- Copilot may suggest irrelevant or incorrect actions
- The AI assistant becomes less helpful and intuitive to use
The label should provide a concise, human-readable name for the action, while the description should explain what the
method accomplishes. These descriptions are crucial for Copilot’s intent matching algorithm to work effectively.
What is the potential impact?
Poor intent matching in Einstein Copilot leads to frustrated users who cannot effectively interact with custom actions through natural language.
This reduces the value of the conversational AI assistant and may result in users avoiding the feature entirely.
