Skip to main content
These are built-in features that facilitate complex, multi-step agency.
- Memory: Memory is essential for maintaining context and learning across task executions.
- User Memory: Stores persistent knowledge about the end-user for personalization.
- Workspace Memory: Stores organizational context shared across agents.
- Agent Memory: Short-term context (previous steps, tool outputs) for the current execution loop.
- Canvas: Enables the Agent to generate structured, shareable digital assets (e.g., text files, documents, spreadsheets).
- Use Case Example: If a user requests a specific segment (e.g., “CROs in San Francisco”), the Agent delivers the output as a fully formatted spreadsheet file.
- WebSearch: Provides real-time data retrieval from the internet.
- Use Case: Critical for tasks requiring current market information (account research) or validating company information that may have recently changed.
- Suggested Action: A proactive workflow assistant that suggests the next best action after providing an output.
- Example: After retrieving a list of contacts (e.g., CROs in San Francisco), it suggests follow-up actions like, “Should we add them to [Sequence XX]?”
