Supply Chain AI Toolkit

What I built

The same three analyses kept recurring across planning, freight and sourcing work, and each time they were rebuilt by hand in a spreadsheet against whatever export was available. This toolkit turns them into reusable browser workflows: load the CSV you already have, map the columns once, and get the calculation back in a form you can filter, sort and export.

Three workflows

Inventory Planner

Calculates days of cover, lead-time demand, reorder points, inventory position and MOQ-rounded order recommendations.

Freight Exceptions

Benchmarks cost per mile by lane and flags high-cost, late and critical shipments.

Supplier Scorecard

Scores suppliers across delivery, quality, cost and service using editable weights.

How it works

Upload CSV → map columns → analyze → filter/sort → export results

You do not need exact header names. Common variants of each field are matched automatically, so a file that calls a column On Hand, on_hand or Qty Available resolves the same way. Anything left unresolved is mapped by hand before the analysis runs, and the mapping is shown so you can see what was matched to what.

Build approach

• The calculations are deterministic. The same file and the same settings produce the same numbers every time — the arithmetic is written out in code, not generated.
• Uploaded CSV processing happens in the browser. No database is required, and nothing is stored between sessions.
• The optional AI Brief writes up results that already exist. It summarizes the calculated output in plain language; it does not perform the calculation and cannot change it.
• Raw uploaded CSV files are not sent to the AI endpoint. Only the calculated results the brief describes are.