What's New in this Release
Mark a Component as AI Agent
With Hero AI, User can now explicitly mark Components as AI agents within Canvas, making it With Hero AI, users can mark components as AI agents in Canvas to identify, organize, and work with AI-powered automation. Mark any component as an AI agent using the "Mark as AI Agent" toggle in the Hero AI section of the Summary tab. Components marked as AI agents appear in the AI Agents list and are available in the AI Agents tab when adding actions to playbooks.
A new AI Agents tab is available under the Orchestration menu to create, update, and view all AI agents. This provides a dedicated space to manage components categorized as AI agents separately from standard components.
For more information, see AI Agents in Orchestrationο»Ώ.
Agentic Hero AI Native Action
Hero AI native action now supports agentic tool use. Orchestrators can configure components as tools for Hero AI native actions in their playbooks. Hero AI analyzes prompts and automatically selects and calls configured tools to gather data and perform actions, enabling it to access additional data sources, services, and systems. This tool-use capability enables Hero AI native actions to handle complex, multi-step prompts that require data gathering and analysis across multiple systems. Tool use improves accuracy and reliability by grounding responses in real-time data from configured components.
For more information, see Hero AI Native Actionο»Ώ.
Import and Export Field Support Update
Previously, list and multi-select fields supported export only. We now support both export and import for list and multi-select fields. The only change in export behavior is that CSV files now use comma delimiters with RFC-compliant escaping, values containing commas are quoted, to ensure correct round-trip import for these field types.
For reference fields, import is now supported. Export will be enabled once reference fields are fully supported in reports. Until then, reference field data can be imported but not exported.
Additionally, a previously released change that caused unintended splitting of list values during record updates has been reverted, restoring the expected behavior.
Auto-Refresh Updates for Dashboards
Auto-refresh behavior has been standardized across dashboards to improve performance and consistency. For dashboards, supported auto-refresh intervals now start at one minute; sub-minute options (such as 5s, 10s, and 30s) have been removed. Existing dashboards configured with unsupported intervals have been automatically adjusted to the nearest valid interval (for example, 30 seconds is updated to 1 minute).
Global Logout
Previously, global logout was controlled through a query parameter that logged users out of all active sessions. Starting in 25.5.0, this behavior is managed through an account-level setting, allowing you to explicitly choose between session-only logout and global logout. As part of this change, the query parameter is deprecated and no longer used.
For more information, see Global Logout Behaviorο»Ώ.