PatternLab Cortex is the operational semantic layer for the shrimp plant. Its ready-made model connects count grades, yield pools, chill clocks, IQF capacity, orders and inventory. The live Supply Chain workspace exposes that model through Plan, Inventory and Sales views.
A quote takes days — someone checks the count mix, the yield and whether the cold chain can even hit the date, by hand. The plan is stale the moment a pond comes in short or a tunnel goes down. And when a buyer or an auditor asks which lots went into a container, it’s a hunt through cartons and clipboards. The big planning suites hand off at the plant door and never touch the floor where the day is actually decided.
AI maps existing ERP fields, receiving records, grading observations and cold-chain documents into Cortex. Your team reviews the mappings and every accepted change remains traceable and versioned. Cortex then derives the pool, validates the constraints, explains the answer and solves one live plan.
An enquiry lands: a grade, a form, a market, a date. SeaTrace checks it against the live plan — is the count in the mix, does the cold chain make the cut-off, what will it cost to hold? It returns a quote you can actually keep, backed by a real slot on the schedule. Automated end to end, from enquiry to quotation — so you stop promising dates the freezer can’t hit.
One pond grades out across ten count bands and splits into co-products at real yield factors — HOSO, HLSO, PD, PUD, cooked. SeaTrace plans the split: it routes the premium counts to the highest-paying orders and absorbs the rest, filling the most valuable book on the least raw shrimp you have to hold. A 26/30 never becomes a U15 — grades route, they don’t remix.
Peel lines, the cook step, the IQF tunnels, blast and holding cold store, the reefer cut-off — finite, sequential, changeover-aware. SeaTrace schedules them together, with hold time and −18°C as hard rules rather than footnotes, so the freezer is never the silent reason a promised date slips.
A pond comes in short. A tunnel drops out. A rush order lands. A lab result holds a lot. SeaTrace re-plans in real time — it re-promises what it can, flags what’s now at risk, and drafts the buy list to cover the gap. The planner doesn’t rebuild the schedule from scratch; they review the change and approve it.
The same connected model that supports the promise also explains its genealogy. Open any container and Cortex walks it back — carton → pallet → IQF batch → peel run → receiving → harvest → pond — with the source records and cold-chain documents linked at each step.
Planners see the feasible day, inventory teams see the live yield pool and cold-store exposure, and sales sees dates backed by the same constraints. Shrimp is a live PatternLab industry model, activated through AI-assisted mapping rather than a blank implementation.