What problem does it solve? Operating Blue Yonder TMS through an AI assistant risks costly mistakes: tendering a load commits carrier capacity and freight spend, approving freight settlement pays money out, and auto-tenders or scheduled optimizer runs can fire commitments with no human in the loop. This Skill gives the model the operator judgment to classify every TMS action as read, reversible write, committing write, or destructive, and to pause consequential ones for human approval. ## Core Features & Use Cases - Operation gating matrix: Classifies Blue Yonder TMS operation families (consolidation, rating, tendering, booking, freight audit and payment) into read / write / destructive tiers with gate rules per tier. - Edge-state and recovery playbooks: Covers multi-leg partial acceptance, spot tender auto-award, self-invoice/auto-pay re-rates, closed finance periods, and non-idempotent commit retries, with recovery patterns for each. - Use Case: A user asks the assistant to tender a load to a carrier. The Skill requires re-reading tender status and carrier eligibility, confirming business unit, lane, mode, and buy cost, and routing the EDI 204 tender through a human approval gate before it fires. ## Quick Start Ask the assistant to rate and tender a Blue Yonder load to a carrier and it will classify each step, show the cost, and pause for approval before committing.
