Intelligent inventory agents for a wholesale distributor

The purchasing team reviewed inventory min and max levels every two to three months instead of monthly, covering only a subset of high priority brands. The majority of inventory never received optimization analysis, leading to both stockouts on fast movers and excess carrying costs on slow movers.

Customer ordering pattern changes went undetected until team members happened to notice a drop in activity. No proactive system existed to flag disengaging customers to sales representatives, which meant retention outreach arrived weeks or months too late.

All analysis happened through manual review in Sage 300 ERP, consuming 2 to 3 hours per cycle. The team lacked the bandwidth to expand coverage, creating a bottleneck that limited the business's ability to respond to market changes.

Built a Customer Retention Agent that pulls order history from Sage 300 via REST API, calculates each customer's normal ordering frequency using statistical analysis, and flags accounts showing deviation patterns. The agent generates personalized retention emails and routes them to the assigned sales representative for review before sending.

Designed a Min Max Agent that analyzes inventory levels, historical ordering patterns, back orders, and movement data across all product brands. The agent produces adjustment recommendations with supporting rationale, then presents them through an admin dashboard where managers approve or decline each suggestion.

Implemented a human in the loop approval system with an admin dashboard. Every recommendation from both agents passes through human review before taking effect. The system logs all approval decisions and overrides, creating a feedback loop that improves recommendation accuracy over time.