Pre-Ship Pick Accuracy Audit

The problem this kills

Wrong items go out the door. A picker grabs the close-but-wrong SKU, a packer slips an extra unit in the box, and the first time anyone notices is when an angry customer emails or files a chargeback. By then you've eaten the return shipping, the replacement, and a chunk of trust.

Most warehouses "audit" by gut feel - a supervisor eyeballs a few boxes when they have a minute, never the same way twice, and nobody writes anything down. So you can't tell which picker is drifting, which SKUs get fat-fingered, or whether last month's coaching actually worked. The errors that slip through aren't random; they're a pattern you can't see because nobody is measuring it.

This plan kills that blind spot. It gives you a disciplined, repeatable re-check on a sample of every batch, a hard hold on anything that fails, and a running scoreboard so problems surface as numbers you can coach against - before the order ships, not after the refund.

What you'll build

A small web app your QC team logs into that:

• Pulls in your packed-and-ready-to-ship orders (from a CSV or Google Sheet to start - no integration needed).
• Randomly selects a configurable percentage of those orders for re-check - and lets you dial the rate up for new or error-prone pickers.
• Walks the auditor through each sampled order, recording a clean pass/fail plus the specific error type (wrong item, wrong quantity, missing item, damaged, wrong address, and your own categories).
• Puts any failed order on hold so it cannot be released until it's corrected and a person re-approves it.
• Keeps a running accuracy scoreboard by picker and by error type, and exports it as a CSV you can drop into a review.

What's inside the Implementation Plan

The plan is a single file you paste into an AI coding agent (Claude Code), and it builds the tool with you step by step. A few things that make it different from a generic template:

• It interviews you first. Before writing a line of code, the plan has the agent ask you about your real picking and packing flow, what your packed-order list looks like, how you name pickers and SKUs, your typical and peak daily volumes, the sampling rules you want, and the messy exceptions you hit. It reflects a short tailored spec back to you for a thumbs-up - so the tool fits your warehouse, not someone's idea of an average one.
• A clear definition of done so you know exactly when you're finished.
• Every build step ends with a ready-to-paste prompt - you copy, paste, and watch it build.
• A "No API yet?" fallback so you can ship the whole thing today using just a spreadsheet, and export results as a CSV in the exact columns your system expects.
• Plain language throughout, with the few unavoidable technical terms explained in a handful of words.

The governance it includes (this is the point)

This isn't a toy. The plan builds in the controls that make a QC tool trustworthy:

• Login so only your team can use it.
• Row-level security so each organization only ever sees its own data.
• A complete audit trail - who audited which order, what they found, and exactly when.
• A hard human-in-the-loop gate: a failed order is held, and only a person can re-approve and release it after correction. The AI never releases anything on its own.
• Duplicate guards so the same order can't be audited or released twice.

Who it's for

QC auditors, fulfillment supervisors, and ops managers who own pick/pack accuracy - anyone who is tired of finding out about errors from the customer instead of from their own process. You do not need to be a developer. If you can use a spreadsheet and follow steps, you can build this.

You've got this - paste the first prompt and let the plan tailor the rest to your warehouse.