OS&D Exception Logger: Clean Evidence for Freight Claims & Chargebacks
Let receiving log over/short/damaged exceptions against the PO with photos and condition notes, match them to expected quantities, and route every record through a supervisor before it triggers a claim or vendor chargeback.
An internal web tool where receiving logs an over/short/damaged exception against a PO line, attaches dock photos and condition notes, the tool compares expected vs received, a supervisor reviews and approves each record, and only then does it produce a clean OS&D record plus a claim / vendor-chargeback trigger and export.
Before you start
- A Supabase account (free)
- A Vercel account (free)
- Exports you already have: a purchase-orders / expected-receipts CSV (PO, line, item, expected qty)
- Your OS&D codes (your over / short / damage reason codes)
- A phone or tablet at the dock for photos
- Claude Code or any AI coding agent
The problem this kills
A truck backs into the dock. The paperwork says 200 cartons; the count comes up 188. Two pallets are crushed. One item showed up that nobody ordered. Your receiver scrawls "12 short, 2 damaged" on the delivery receipt, maybe snaps a photo on a personal phone, and the day moves on.
Then, weeks later, you try to file a freight claim or charge the vendor back — and the evidence has evaporated. The photo is on someone's phone that left the company. Nobody wrote which PO line was short. The condition note says "damaged" with no detail. The carrier denies it, the vendor disputes it, and money you were owed quietly disappears. Over, Short & Damaged (OS&D) exceptions are real cash, and they only get recovered if the proof is captured cleanly at the moment of receipt — by the person on the dock, the instant they see it.
You don't need a warehouse management suite to fix this. You can build the logger yourself, in an afternoon.
What you'll build
An internal web tool your receiving team opens on a phone, tablet, or dock computer. When a shipment doesn't match the paperwork, the receiver picks the PO, picks the line, and logs the exception: how many were expected, how many actually arrived, an over / short / damage code, a plain condition note, and photos snapped right there at the dock. The tool does the math — expected vs received — and labels each exception so there's no guesswork later.
Every logged exception waits in a review queue. A supervisor opens it, sees the photos and the variance, and approves or rejects. Only on approval does the record become an official OS&D record and fire a claim or vendor-chargeback trigger — with all the evidence attached. A duplicate guard (PO + line + receipt date) stops the same shortage from being logged twice. And there's a clean CSV export in the exact columns your claims system or vendor-scorecard spreadsheet expects, so this is fully usable today with no integration at all.
What's inside the Implementation Plan
The downloadable plan is a single markdown file you paste into an AI coding agent. It opens by interviewing you about your own receiving process — your PO and line naming, your real over/short/damage codes, how your delivery receipts and expected-quantity files look, who's allowed to approve a chargeback, and your messy edge cases (substitutions, partial deliveries, mixed pallets) — and then it reads a short spec back for your thumbs-up before it builds anything. That's the difference between a tool shaped to your dock and a generic template you have to fight.
From there it walks the agent through the data model (POs and expected lines, exceptions, photos, review decisions), the photo capture and file Storage, the expected-vs-received math, the duplicate guard, the supervisor review-and-approve screen, the claim/chargeback trigger, and the CSV export. Every step ends with a ready-to-copy prompt. There's a full "No API yet?" path: import expected quantities from a CSV, export a clean OS&D CSV, and you never have to touch your WMS or ERP API to ship.
The governance it includes (this is the point)
OS&D feeds money decisions — claims and chargebacks — so the controls are the product. The plan builds them in: a login so only your team can use it; row-level security so each site or organization only ever sees its own exceptions; a complete audit trail of who logged, edited, approved, rejected, and exported, and when; a hard human-approval gate so no claim or chargeback can fire until a supervisor signs off on the photos and the variance; and a duplicate guard (PO + line + receipt date) so the same shortage can't be logged or charged back twice. An exception physically cannot become a claim trigger until a person with authority approves it. That's the evidence trail a carrier or vendor can't wave away.
Who it's for
Receiving and warehouse staff who log what actually showed up, and the claims or logistics supervisors who turn those exceptions into recovered money. If you can explain to a new dock hire what counts as "short" versus "damaged" and who has to sign off before you bill the carrier, you can build this — no developer required.
You've got this — open the plan, paste the first prompt, answer a few questions about how your dock actually runs, and you'll log your first clean exception with photos this afternoon.