Supplier Quality Defect Tracker
Pull together every inspection failure, NCR, and defective return by supplier, let AI compute a volume-weighted defect-rate / PPM scorecard and rank your chronic offenders, then draft a quality alert or debit-memo request — sent only after a quality manager approves.
A web tool where you import inspection, NCR, and return defect history plus units received per supplier; AI normalizes defect categories across those sources, dedupes on the defect event, computes a defect rate and PPM per supplier weighted by volume, attaches the cost of poor quality, and ranks the worst offenders. A quality manager reviews the scorecard, approves it, and only then does the tool email the quality scorecard and export a CSV debit-memo list for the chronic suppliers.
Before you start
- A Supabase account (free)
- A Vercel account (free)
- A Resend account (free)
- Your defect history (inspection failures, NCRs, returns) and units-received-per-supplier as CSV or a Google Sheet
- Claude Code or any AI coding agent
The problem this kills
Your defect data is scattered. Incoming inspection failures live in one spreadsheet. Non-conformance reports (NCRs) sit in a binder or a different tab. Defective customer returns and field complaints land somewhere else entirely. Each source names the same problem differently — "dent" here, "physical damage" there, "cosmetic" in a third place — and none of them know how many units you actually received from that supplier, so you can't tell whether a vendor with twelve defects is worse than one with four.
So when it's time to score suppliers or push back on a chronic offender, you spend a day stitching it all together by hand: copy-pasting between sheets, guessing which categories mean the same thing, dividing defects by received units to get a rate, and trying to remember which defective event you already counted from two different reports. The math is easy to get wrong, the same return shows up twice, and the supplier rightly argues with a number you can't defend. The chronic problem suppliers keep shipping, and the cost of poor quality quietly eats your margin.
This is exactly the kind of rules-based, multi-source roll-up that a small internal tool does far better than a spreadsheet — and you do not need to be a developer to build it.
What you'll build
A simple internal web tool for supplier quality. You import your defect history from all three sources — incoming inspection failures, NCRs, and defective returns — plus a file of units received per supplier for the period. The tool normalizes the defect categories so the same problem from different sources counts as one thing, dedupes on the defect event so nothing is counted twice, and computes a per-supplier scorecard: total defects, defect rate and PPM (defects per million units received) weighted by volume, and the cost of poor quality when you've loaded cost data. It then ranks your worst offenders against the thresholds you set.
The quality manager reviews the whole scorecard on one screen, adjusts or excludes anything that looks off, and clicks Approve. Only then does the tool email the finished quality scorecard to whoever you choose and export a clean CSV debit-memo request list for the chronic suppliers — in the exact columns your purchasing or ERP system expects.
What's inside the Implementation Plan
The plan is a single file you paste into an AI coding agent. It opens by interviewing you about your business — which systems hold your inspection, NCR, and return data, the exact columns and the defect-category wording each one uses, how you identify one supplier and one defect event, your typical and peak volumes, and the exact rules that make a supplier a "chronic offender" — and then tailors the data model, the category normalization, the PPM math, and every later step to your answers. This is a build shaped around your quality process, not a generic template.
From there it walks the agent through the database schema, the multi-source CSV imports with their duplicate guards, the category-normalization and volume-weighted scoring engine, the quality manager's review-and-approve screen, and the export of the scorecard email and the debit-memo CSV. Every step ends with a ready-to-copy prompt. Because the whole thing runs on CSV in and CSV out, you can build and use it this weekend even if nothing you own has an API.
The governance it includes (this is the point)
This drives money and supplier relationships, so it's built like it matters: login so only your quality and purchasing team can use it, row-level security so you only ever see your own organization's data, and a complete audit trail of every import, edit, approval, and send — who did what, and when. Nothing leaves the building automatically: the scorecard and any debit-memo request are a draft until the quality manager approves them, and approval is the hard human-in-the-loop gate before a single email or debit memo goes to a supplier. Duplicate guards on the defect event mean the same inspection failure or return can't inflate a supplier's count twice.
Who it's for
Supplier-quality engineers, purchasing and sourcing teams, and quality managers who own supplier performance and want a defect scorecard that's fast, volume-fair, and defensible enough to put in front of a vendor. If you can explain how you decide a supplier is a problem, you can build this.
You've got this — open the plan, paste the first prompt, and let it interview you about your quality process.