Machines Triage. Humans Decide.

This analysis recommends designing SOC workflows that assign deterministic, low‑risk triage tasks to AI agents while reserving ambiguous or asymmetric, high‑impact decisions for human analysts; it emphasizes four architectural requirements—explainable model verdicts, explicit boundaries for autonomous actions, analyst feedback loops, and complete audit trails—to maintain accountability and operational effectiveness.