Prompt Chain: Weekly Inventory & Restock Decision System

What This Builds

A weekly 3-step prompt chain that takes raw inventory data from your Seller Central export, runs it through ChatGPT in sequence, and produces: (1) a prioritized restock list with urgency levels, (2) draft purchase orders for each supplier, and (3) a management summary you can send in 5 minutes. What takes 2-3 hours of manual spreadsheet work on Monday mornings becomes a 20-minute process.

Prerequisites

• ChatGPT account ({{tool:ChatGPT.plan}} at {{tool:ChatGPT.price}} recommended. larger context window handles more SKUs)
• Google Sheets with your inventory data (or ability to export from Amazon Seller Central)
• Supplier names, lead times, and minimum order quantities for your main SKUs
• Time to build: 1.5 hours; weekly use takes 20 minutes
• Cost: {{tool:ChatGPT.price}}

The Concept

A prompt chain is a relay of AI instructions where each step's output feeds into the next step's input. Think of it as an assembly line: the first station takes raw data and sorts it, the second station turns the sorted data into purchase orders, and the third station writes the executive summary. You couldn't get a good executive summary from raw data in one step, but breaking it into stages produces professional results.

Build It Step by Step

Part 1: Prepare Your Data (15 minutes)

Step 1: Export inventory data from Amazon Seller Central

1. Log into Seller Central
2. Go to Inventory → Manage FBA Inventory → Download Inventory
3. Export as CSV. Open in Google Sheets.
4. You need these columns minimum: Product Name, ASIN, Current Units (FBA), Units Sold Last 30 Days, Inbound Units (in transit)

Step 2: Add supplier data to your sheet

1. Add two columns manually (this only needs to be done once):

• Supplier Name, which vendor supplies this product
• Lead Time (Days). how many days from order to FBA-ready

1. Add a calculated column in the sheet (use ChatGPT to build this formula if needed):

• Days of Stock = Current Units ÷ (Units Sold Last 30 Days ÷ 30)

Step 3: Export as plain text for ChatGPT

1. Select your data range and copy it (Ctrl+C / Cmd+C)

Part 2: Step 1 of the Prompt Chain. Triage

1. Open ChatGPT ({{tool:ChatGPT.url}}) and start a new conversation with {{tool:ChatGPT.model:reasoning}}
2. Paste this prompt, then paste your inventory data below it:

You are my inventory analyst. Below is my current inventory data with days-of-stock estimates.

Categorize every SKU into:
🔴 CRITICAL: under 14 days of stock (reorder immediately)
🟡 PRIORITY: 14-30 days of stock (reorder this week)
🟢 OK: 30+ days of stock (no action needed)

For each CRITICAL or PRIORITY item, calculate the recommended order quantity using this formula:
(60 day supply at current velocity) - current stock - inbound stock

Output as a clean table with columns: Product Name | Days of Stock | Status | Recommended Order Qty

Here is the inventory data:
[PASTE YOUR DATA]

1. Save the output. copy it to a new section of your Google Sheet or a text file.

Part 3: Step 2. Generate Purchase Orders

1. With the triage output still in the same ChatGPT conversation, add:

Great. Now for every CRITICAL and PRIORITY item, group them by supplier and create a draft purchase order for each supplier.

Format each PO as:
---
TO: [Supplier Name]
DATE: [today's date]
SUBJECT: Purchase Order Request

Hi [Supplier],

We need to reorder the following items. Please confirm availability and estimated ship date.

[Product | SKU/ASIN | Units Requested | Notes]

We need these to arrive at our warehouse by [date 30 days from today].

Please confirm receipt and provide a proforma invoice.

Thank you,
[My Name]
[Company]
---

1. Copy each purchase order draft. You'll paste these into email to send to your suppliers.

Part 4: Step 3. Management Summary

1. Still in the same conversation, add this final step:

Finally, write a 5-bullet management summary of this week's inventory situation. Include:
- How many SKUs are in CRITICAL status and what's the risk
- How many are in PRIORITY status
- Total approximate reorder value (estimate based on average $X per unit if I haven't given you pricing)
- Any patterns you notice (e.g., multiple products from one supplier all running low)
- Recommended next action

Keep it concise — my manager reads this in 60 seconds.

Real Example

Setup: 45-SKU catalog. Monday morning. You have 2 hours of inventory work ahead.

Step 1 input: Paste 45 rows of inventory data with the triage prompt. Step 1 output: A clean table. 4 items are CRITICAL (under 14 days), 7 are PRIORITY, 34 are OK. The system flags that your best-selling item has only 9 days left.

Step 2 input: "Generate purchase orders" (same conversation). Step 2 output: 3 purchase orders (one per supplier who needs restocking) with exact units requested, formatted and ready to email.

Step 3 input: "Write the management summary." Step 3 output: "4 SKUs require immediate restock action, including our top-selling item with 9 days of stock remaining. 7 additional SKUs require orders placed by Thursday to maintain 30-day buffer. All critical items sourced from Supplier A, recommend confirming capacity on Monday's call. Total estimated reorder value: ~$8,500."

Total time: 20 minutes vs. 2.5 hours manually.

What to Do When It Breaks

• ChatGPT miscalculates days-of-stock → Double-check your formula in Google Sheets before pasting. ChatGPT is good at analysis but bad at arithmetic, calculate numbers in the spreadsheet first.
• Outputs mix up supplier names → Add a note to Step 2: "Supplier data is in the last column of the inventory table, use exactly the name shown."
• Summary feels generic → Add your specific context: "Our Q4 is our peak season, so 30-day buffer should be 45 days in October through December."
• ChatGPT conversation gets too long → Start a fresh conversation for Step 3. Paste just the Step 2 output and ask for the summary.

Variations

• Simpler version: Only do Step 1 (triage). A categorized list by urgency is already 80% of the value, you draft the POs manually.
• Extended version: Connect Google Sheets to Zapier to auto-trigger this analysis every Monday. The output goes to a Slack channel or email automatically, your inventory report writes itself.

What to Do Next

• This week: Run through this chain manually once with your current inventory data.
• This month: Build the Google Sheets template with all formulas pre-built, weekly use becomes copy-paste.
• Advanced: Connect Zapier to Amazon's inventory API to pull data automatically. The chain triggers every Monday at 7am, the triage lands in your Slack, and you never have to manually export again.

Advanced guide for e-commerce coordinator professionals. These techniques use more sophisticated AI features that may require paid subscriptions.