As of 13 August 2026, AI can only partly set the right stock levels for your business.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Hand it to a person

   Second choice

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

3. Do it yourself

   The distant third

   A chat interface, power-user skill, and roughly 2 hours until you can act on the result.

   How to actually do it

   1. Open your stock, sales and purchasing records and export dated SKU-level data showing units sold, current stock, open orders, stockout periods and returns.
   2. Gather supplier information for each SKU, including usual and worst-case lead times, minimum order quantities, pack sizes, delivery days and ordering frequency.
   3. Write down the constraints that the model cannot infer reliably, including shelf life, storage capacity, cash limits, planned promotions, substitutions and the operational consequence of a stockout.
   4. Paste the prompt and the gathered information into a chatbot, then ask it to produce the table, formulas, missing-data list and low, medium and high protection scenarios.
   5. Compare every input in the response with the live stock system, supplier terms and the date range of your sales data, correcting any stale or misclassified figures.
   6. Recalculate a sample of the demand, safety-stock, reorder-point and order-quantity rows in a spreadsheet and investigate any difference before using the recommendations.
   7. Run the proposed levels through a colleague who knows purchasing and operations, then approve a limited change and monitor stockouts, excess stock and supplier performance before applying it across the business.

   Prompt

   Set recommended stock levels for the products in the data below. This is for a UK business, and the recommendations must be based only on the information I provide. Do not invent demand, lead times, costs, service targets or supplier behaviour. If a required input is missing, identify it and show how the recommendation would change under clearly labelled assumptions instead of filling the gap silently.
   
   Use these definitions where the data supports them:
   - average demand over the stated period
   - demand variability
   - supplier lead time and lead-time variability
   - safety stock
   - reorder point
   - target stock level or order-up-to level
   - minimum order quantity, pack size and supplier constraints
   
   For each SKU, return a table containing: SKU, product name, current stock, average demand, demand period, supplier lead time, demand during lead time, safety stock, reorder point, recommended maximum or order-up-to level, suggested order quantity, assumptions, missing data and confidence. Explain the formula used for every calculated field. Keep units consistent and round only where the pack size requires it.
   
   Use the service level or stockout tolerance below if supplied. If it is not supplied, do not choose one silently. Provide scenarios for low, medium and high protection and explain the operational and cash implications of each. Flag products with intermittent demand, short shelf life, seasonal demand, promotions, substitutions, uncertain lead times or unusual recent sales rather than treating them as ordinary demand.
   
   After the table, list the three decisions I must make before using the recommendations, the data that should be refreshed before the next review, and a simple monitoring plan for stockouts, excess stock and forecast error.
   
   Business context:
   [describe the business, products, storage limits and consequences of stockouts]
   
   Stock and sales data:
   [paste dated SKU-level stock and sales data, including units and the period covered]
   
   Supplier data:
   [paste supplier, lead time, lead-time range, minimum order quantity, pack size and ordering frequency]
   
   Constraints and targets:
   [paste shelf life, storage limits, service targets, budget limits, promotions and known changes]
   
   Do not present the result as certain or ready for automatic ordering. Finish with a short list of checks I must complete against the live inventory and supplier information before approving any change.

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