Customer Part-Number Cross-Reference Resolver

The problem this kills

Every customer calls the same product something different. Their PO says AB-1140-X, your system calls it SKU-44821, and someone on your team has to know - or look up - that they're the same thing. On a 30-line order that's 30 lookups, and one slip enters the wrong SKU, ships the wrong part, and triggers a return.

Worse, all that matching knowledge lives in people's heads or in a tangle of personal spreadsheets. When the person who "just knows" is out, order entry slows to a crawl, and new hires take weeks to get up to speed.

This tool turns that tribal knowledge into a shared, growing crosswalk. Paste the order, and the parts you've seen before resolve automatically. The ones you haven't get suggestions and a clear "needs review" flag - so nothing wrong ever sneaks through, and every confirmed match makes the next order faster.

What you'll build

A clean internal web app where your team:

• Pastes or uploads incoming order lines - customer ID plus their part numbers and quantities.
• Watches parts auto-resolve to your internal SKUs using your existing crosswalk.
• Sees fuzzy-match suggestions for near-misses (a typo, a dropped dash, an extra digit) - ranked by confidence, never auto-applied when confidence is low.
• Gets unknowns flagged clearly, so no line is silently guessed.
• Routes new and uncertain matches to a reviewer, who approves or corrects them before they're saved.
• Builds the crosswalk automatically - every approved match is remembered, so the tool gets smarter with every order.
• Exports a clean, mapped order in the exact columns your order system expects.

What's inside the Implementation Plan

The plan is a single file you paste into an AI coding agent (Claude Code), which then builds the tool with you step by step - in plain language, no coding background needed.

It opens by interviewing you about your business - your real part-number formats, how your customers label things, your typical and peak order volumes, what counts as a "confident" match, and your messiest edge cases. It reads back a short tailored spec for your thumbs-up, then shapes the database, the matching rules, and every later step around your answers - so you get a tool fit to your operation, not a generic template you have to bend to.

From there it walks through: setting up the database and login, importing your starting crosswalk, building the paste-and-resolve screen, adding fuzzy matching with confidence scores, building the reviewer approval queue, wiring up email notifications, and producing the clean CSV export. Every build step ends with a ready-to-copy prompt.

The governance it includes (this is the point)

This isn't a loose script that guesses and writes. Governance is baked in:

• Login so only your team can use the tool.
• Row-level security so each organization only ever sees its own crosswalk and orders.
• A hard human-in-the-loop gate - the AI drafts matches, a reviewer approves, and only then is anything saved to the crosswalk or applied to the order. Low-confidence matches can never auto-save.
• A complete audit trail - who matched what, who approved it, and when.
• Duplicate guards - the same customer part number for the same customer can't create conflicting crosswalk entries (dedupe key: customer ID + customer part number).

Who it's for

Order-entry, inside sales, and customer service teams who key in orders that arrive in the customer's part-number language - especially distributors, manufacturers, and wholesalers whose customers each have their own naming for the same products.

You don't need to be a developer. If you can fill out a form and copy-paste, you can build this.

You've got this - paste the first prompt and let's build it.