AI Supply Chain & Logistics Agent

Description

We import from eleven countries and our lead times are, politely, aspirational. The planning team works from a supplier-quoted lead time that has been wrong for three years, and everybody compensates with safety stock nobody has recalculated since. We want an agent that builds actual lead-time distributions from our own purchase order history — per supplier, per port, per season — and flags in-flight orders that are tracking late against their own distribution rather than against the quoted date. The second half is safety stock. Given real variability, tell us where we are carrying too much and, more importantly, where we are carrying too little and have been lucky.

• Build empirical lead-time distributions per supplier, route and season from PO history
• Flag in-flight orders tracking late against their own distribution, with an expected arrival range
• Recalculate safety stock against observed variability and target service levels
• Call out the under-stocked positions explicitly, not just the over-stocked ones
• Deliver into the planning team's existing tooling rather than a separate dashboard

• Lead-time distribution model from purchase order history
• In-flight order risk flagging with arrival ranges
• Safety stock recommendations, over and under, with service-level assumptions stated
• Integration into the planning team's existing tools