Stuck-Order Escalation Ladder

The problem this kills

Orders rarely fail loudly. They stall. One sits in "Awaiting Picking" for nine days because a SKU was short. Another is stuck in "Pending Credit Approval" while everyone assumes someone else is on it. By the time the customer calls, the order is a fire - and nobody owned it the whole time it was burning.

The real problem isn't that orders get stuck. It's that nothing watches the clock per stage, and nobody is automatically put on the hook when a stage runs long. Your team is heroic at firefighting and terrible at being warned before the fire. Spreadsheets of open orders don't help, because they don't know how long is too long for each specific stage.

This tool fixes the accountability gap. It knows your per-stage SLA (how many hours or days an order is allowed to sit in each stage), it ranks the orders that have blown past it, and it walks each one up a defined ladder of owners - CSR, then team lead, then ops manager - escalating one rung at a time until someone acts. And because escalations get noisy fast, a manager reviews and approves the list before a single alert fires.

What you'll build

A small, login-protected web app that:

• Ingests your order stage history - either from a CSV / Google Sheet you upload, or (later) a live feed - so it knows which stage each order is in and how long it's been there.
• Compares each order against its per-stage SLA and flags the ones over the line.
• Ranks the breaches so the worst (most overdue, highest value, oldest) float to the top.
• Maps each breach to a rung on your escalation ladder and figures out who should be alerted next.
• Puts a human in front of the trigger: a manager sees the proposed escalation list, can drop or hold individual rows, and only then approves.
• Fires alerts on approval - emails the right owner via Resend, records the new owner, and logs everything.
• Keeps a complete audit trail - who approved what, who got alerted, and when - and never escalates the same order at the same stage twice thanks to a duplicate guard.

What's inside the Implementation Plan

• It starts by interviewing you about your business. Before it writes any code, the plan has the AI agent ask you a tailored set of questions - your real stages and their names, your SLA for each, your escalation ladder and who sits on each rung, your typical and peak order volumes, and your messy exceptions (rush orders, holds, backorders). This is not a generic template; the data model and rules are shaped around your answers, and the agent reads a short spec back to you for a thumbs-up before building.
• A clean, copy-paste prompt for every build step, in order.
• A working data model for orders, stage history, SLAs, your escalation ladder, and the audit log.
• The ranking logic that decides which stuck orders matter most.
• The manager approval screen - the human gate before anything fires.
• Resend-powered escalation emails with the right owner on the right rung.
• A "No API yet?" fallback so you can build the whole thing today from a CSV / Google Sheet and export a clean CSV in the exact columns your system expects.
• A verification checklist so you know it actually works.

The governance it includes (this is the point)

This is an internal tool that changes who is accountable for an order, so it's built to be trustworthy from day one:

• Login so only your team can open it.
• Row-level security so each organization only ever sees its own orders and ladder - no leaks across tenants.
• A hard human-in-the-loop approval gate: the AI drafts the escalation list, a manager reviews and approves, and only then do alerts fire and owners change. Nothing escalates on autopilot.
• A complete audit trail: who approved, who was alerted, which rung, and exactly when.
• A duplicate guard keyed on order ID + stage so the same stuck order can't be escalated twice for the same stall - no double-pinging your team.

Who it's for

Ops managers, CSR (customer-service) leads, and fulfillment teams who are tired of orders quietly dying in a stage with no owner. If you live in your order status reports and keep discovering stalls after the customer complains, this is your early-warning system - and your accountability ladder.

You've got this. Open the plan, paste the first prompt, and let the agent interview you.