As of 13 August 2026, AI can only partly create a stock replenishment plan.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Do it yourself

   Second choice

   A chat interface, chat-fluent skill, and roughly 1 hour until you can act on the result.

   How to actually do it

   1. Open your stock system and export the current stock, stock on order and SKU list for the planning period.
   2. Gather recent sales by SKU, supplier lead times, minimum order quantities, case sizes, unit costs, shelf-life limits and any known promotions or seasonal changes.
   3. Paste the business rules and the product data into the prompt, keeping the column names and units clear and marking unavailable fields as missing.
   4. Ask the model to produce the replenishment table, calculations, exceptions, assumptions and approval checklist specified in the prompt.
   5. Compare every current-stock, on-order, sales, lead-time, cost and supplier figure in the draft against the stock system, purchase records and current supplier information.
   6. Recalculate the proposed order quantities and arrival dates for unusual or high-value items, then amend the plan for promotions, expiry risk, supplier uncertainty and minimum order rules.
   7. Send the checked plan to the person responsible for purchasing or operations for approval, and place orders only through your normal purchasing process.

   Prompt

   Create a stock replenishment plan from the data below. Use only the figures and rules I provide. Do not invent demand, stock levels, lead times, supplier terms, minimum order quantities, pack sizes, shelf lives or prices. If a required field is missing, mark it as missing and show what decision it prevents.
   
   Business context:
   - Planning period: [dates]
   - Review frequency: [daily, weekly or other]
   - Service level or stockout policy: [rule]
   - Available budget or ordering constraint: [rule]
   
   Product data:
   [Paste one row per SKU with: SKU, product name, current stock, stock already on order, recent sales by week or day, forecast or expected demand if available, supplier lead time, minimum order quantity, order multiple or case size, unit cost, shelf life or expiry constraint, storage constraint, and supplier.] 
   
   Required output:
   1. A table with SKU, current available stock, demand used, lead time, reorder point, recommended order quantity, suggested order date, expected arrival date, and the reason for each recommendation.
   2. Separate urgent orders, routine orders and items that should not be ordered.
   3. Show the calculation for each recommendation in plain language.
   4. Identify duplicate, inconsistent or stale-looking data without correcting it silently.
   5. List assumptions and missing inputs separately.
   6. Flag items where a human must decide because of promotions, seasonal demand, new products, supplier uncertainty, expiry risk or unusual sales.
   7. Do not place an order or claim that the plan is optimal. End with a short checklist of records I must verify before approval.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

3. Hand it to a person

   The distant third

   A person who owns the outcome does this end to end, worth it when the failure is dear.