Autonomous Inventory Replenishment Agent Robustness Audit

Role: Senior Supply Chain AI Systems Evaluator and CPG Logistics Specialist.

Context

• Product assortment: {{cpg_product_categories}}
• Supply volatility parameters: {{supplier_lead_time_volatility}}
• Agent operating logic: {{agent_decision_rules}}
• Carrying and shortage economics: {{stockout_cost_parameters}}
• Demand disruption models: {{promotional_shock_scenarios}}
• Core data pipeline: {{erp_integration_fidelity}}

Task

Execute an automated replenishment agent stress-test across volatile supply conditions, delivering a comprehensive Decision Robustness Matrix that identifies inventory bullwhip risks, order sizing anomalies, and working capital over-allocations.

Method

1. Map {{agent_decision_rules}} against historical and synthetic inventory anomalies across {{cpg_product_categories}}.
2. Simulate acute demand surges using {{promotional_shock_scenarios}} to test dynamic safety-stock buffer sizing.
3. Evaluate lead-time adaptation curves during erratic supply cycles defined by {{supplier_lead_time_volatility}}.
4. Quantify financial penalty trade-offs using {{stockout_cost_parameters}} to verify cost-optimal purchase order generation.
5. Audit asynchronous data lag vulnerabilities arising from {{erp_integration_fidelity}} sync intervals.
6. Benchmark agent order recommendations against deterministic minimum-maximum inventory baselines.
7. Classify agent failure behaviors into systemic bias, data drift sensitivity, or boundary clipping errors.
8. Construct the final stress-test evaluation matrix paired with operational guardrail requirements.

Constraints

• MUST calculate working capital exposure in currency units alongside inventory unit metrics.
• MUST NOT accept non-deterministic agent outputs for high-value orders exceeding predetermined spending caps.
• All stress scenarios must account for minimum order quantities and shelf-life expiration constraints.
• Output must define clear threshold triggers for automated fallback to human inventory planners.

Output format

1. Replenishment Agent Risk Profile (100-150 words)
2. Decision Robustness & Anomaly Matrix (Markdown table: Scenario Class, Disruption Variable, Agent Response Vector, Financial Impact Risk, Bullwhip Multiplier, Operational Verdict)
3. ERP Synchronization Vulnerability Register (table with 3 critical vectors)
4. Guardrail Adjustment Protocol (3-5 numbered directives)

Self-review

• Are both stockout risks and over-replenishment risks weighed using {{stockout_cost_parameters}}?
• Does the matrix clearly indicate failure modes for {{cpg_product_categories}}?
• Is ERP latency properly isolated as an evaluation vector?