From zero-shot forecast to purchase order with Amazon Bedrock AgentCore

Summary

This post presents an inventory automation architecture that moves from zero-shot demand forecasts to validated purchase orders while applying safety stock, minimum order quantities, budgets, and promotional adjustments. It combines Amazon Chronos2, hosted on SageMaker Serverless Inference, with four Claude-powered agents built using the Strands Agents SDK and deployed on Amazon Bedrock AgentCore; deterministic Python tools perform computation and agents handle judgment. Product data and future covariates arrive through CSV files in S3, while business rules remain in JSON, enabling what-if forecasts and onboarding without per-SKU training. In internal testing across 50 SKUs over a four-week horizon, the system reached 12.3% median WAPE, reduced onboarding from 2–3 weeks to under five minutes, and lowered estimated monthly inference cost from about $1,091 to $15. A conditional retry loop surfaces budget violations for user decisions instead of silently truncating orders, with traces and evaluator scores supporting auditability.

Takeaways

1. 1

   Chronos2 accepts historical and future covariates through the predict_df() API, enabling promotion and price what-if scenarios without per-SKU fine-tuning.

2. 2

   In internal testing across 50 SKUs over a four-week horizon, the system achieved 12.3% median WAPE and averaged 8 seconds of end-to-end latency per SKU, excluding cold start.

3. 3

   When a budget constraint would cause a shortfall, the Supervisor presents options such as raising the budget, expediting delivery, or delaying a promotion instead of silently truncating the order.