Messy Order CSV Normalizer: One Format for Every Customer's Spreadsheet

The problem this kills

Every customer sends their orders a little differently. One emails an Excel file with merged header cells and quantities in cases; another sends a CSV where the SKU is buried in a description string; a third lists pallets, not eaches, and uses their own part numbers that don't match yours at all. Someone on your team opens each file, squints at the columns, mentally converts cases to eaches, hunts down which of your SKUs each line really means, and re-keys the whole thing into your ERP — for every order, every day.

It's slow, it's error-prone, and the mistakes ship: a unit-conversion slip that triples a quantity, a SKU match that's almost-but-not-quite right, the same PO keyed in twice. Onboarding a new customer's file format becomes its own little project. You don't need to live like this, and you don't need to be a developer to fix it.

What you'll build

A simple internal web tool. You upload a customer's order file — any layout, any units, any SKU convention — and the tool reads it, then AI proposes a three-part mapping: which incoming columns become your standard fields, how each incoming unit of measure (case, pack, pallet, each) converts to your base unit, and which of your master SKUs each incoming item code or description matches. A reviewer sees the proposed mapping side by side with the raw file, fixes anything wrong, and clicks Confirm. The confirmed mapping is saved as a reusable template for that customer, so the next file from them maps itself in one click. On confirm, the tool produces a clean normalized order and an ERP-ready CSV in exactly the columns your system expects.

What's inside the Implementation Plan

The downloadable plan is a step-by-step file you paste into an AI coding agent. It opens by interviewing you about your business — who handles order entry today, which ERP you key into and its exact import columns, how your customers' files actually look (sample headers and rows), your real SKU and unit-of-measure conventions, your case/pack/pallet ratios, your typical and peak order volumes, and the messy edge cases that bite you — and then it tailors the data model, the mapping logic, the unit-conversion math, and every later step to your answers. This is not a generic template; the agent reflects a short spec back to you and waits for your thumbs-up before it builds anything. From there it walks the agent through the file upload and parsing, the AI mapping proposal, the per-customer template save/reuse, the reviewer confirm gate, the unit-conversion and SKU-match logic, the normalized order, and the ERP CSV export — each step with a ready-to-copy prompt. There's also a fallback so you can build and use the whole thing today with plain CSV/Sheets in and out, no ERP integration required.

The governance it includes (this is the point)

This handles real orders, so it ships with the controls an operations team needs: login so only your team can use it, row-level security so you only ever see your own organization's customers and orders, a complete audit trail of who mapped, confirmed, and exported which order and when, a hard human-confirm gate so no normalized order or ERP CSV is produced until a reviewer signs off on the column mapping and the line items (especially the unit conversions and SKU matches), and duplicate guards keyed on customer ID + customer PO number so the same order can't be processed twice.

Who it's for

Order-entry clerks, EDI and customer-onboarding specialists, and CSRs who receive order spreadsheets in dozens of layouts and re-key them by hand. If you can describe how a customer's file looks and how it maps to your ERP, you can build this.

You've got this — start with the plan, paste the first prompt, answer the interview, and you'll watch a messy order file turn into a clean ERP-ready order the same afternoon.