Reasoning level attribute

• Type: string
• Value set: Enum
• Data types: Automation
• Fields: reasoning_level

For automations, reasoning_level controls how much deliberate reasoning GovTribe AI uses during a run and its tool work, including a Run now test. The value is independent of a particular AI model.

Values

• none: Use for mechanical scanning, extraction, or explicit rules with no meaningful judgment.
• low: Use for routine classification, summarization, and straightforward fit checks.
• medium: Use for multi-signal comparison, ambiguity resolution, or consequential recommendations. This is the default and recommended starting point for new automations.
• high: Use for difficult multi-step analysis, competing evidence, or substantial planning.

Existing automations keep their saved reasoning level. Create requests that omit the field use medium, while update requests that omit the field preserve the saved value. A later message sent manually in the generated conversation uses that conversation's model and reasoning settings plus the user's current verbosity preference.

Related articles

• Automations (beta): Choose and manage an automation's reasoning level.
• Create automation MCP tool: Set the value when creating an automation.
• Update automation MCP tool: Change the value with a version-gated patch.