Decision Graph learns from your organization's history of human approval decisions. It turns them into Decision Event Records: structured, precedent-based decision intelligence that captures the conditions behind each decision, the outcome, and a confidence score. Built for platform teams, process owners, and approvers who want faster, more consistent, and explainable decisions, it surfaces this intelligence as a subflow tool that Agents can act on or use to justify a recommendation. 

- **Decision Event Records:** Mine past decisions(Approval decisions for V1 release) into conditions, outcomes, and a confidence score.
- **Confidence scoring:** Rank decisions by strength of precedent to set auto-action vs. human-review thresholds.
- **Mine Historical data:** Populate DERs from data that already exists on the instance
- **Subflow-as-a-tool:** Plug into AI agent flows — works alongside your existing approval logic.