You can use AI for inventory management without buying new software. Start with the spreadsheet you already have, add sales history and supplier lead times, and ask an AI chatbot to flag what to reorder, what is moving slowly and what will run out first. Check every number against your sheet before you order. Used this way, AI saves hours and helps a small shop avoid stockouts and dead stock.
What AI actually does for a small stockroom
AI does not count your shelves. It reads the numbers you give it and spots patterns faster than a person scanning hundreds of rows. It can do four jobs: sort products into fast, steady and slow sellers; estimate when each product will run out; draft a suggested order list with quantities and costs; and summarise weekly stock activity.
It cannot know about damaged goods or miscounts, so AI suggests and you decide. Never order on an unchecked AI output.
Start with a spreadsheet, not new software
Most guides jump straight to paid inventory platforms. That is a poor start for a shop with under a few hundred products or a single stockroom. A spreadsheet plus a free chatbot covers the core jobs.
Set up one sheet with these columns: product name, SKU, units on hand, units on order, average weekly sales, supplier name, lead time in days, minimum order quantity, unit cost and selling price. Export sales from your till or online store once a week. Keep one row per product and one consistent unit.
Then follow a fixed weekly routine. On the same day each week, update stock figures, add new sales and run the prompts below. Review only items flagged as order now or watch. This takes most small shops under an hour once the sheet is set up.
If supplier invoices are still paper or photos, scan them first so costs enter the sheet accurately. Our guide to the best AI receipt scanner apps for UK small businesses explains VAT receipt handling.
Work out your reorder point with a simple formula
A reorder point answers one question only: at what stock level should you start ordering? It does not tell you how much to order. Keep those decisions separate, or you will order the right product at the wrong time or in the wrong quantity.
The formula is: reorder point equals average daily demand multiplied by lead time in days, plus safety stock.
Worked example: a gift shop sells 8 candles per day, and its supplier takes 7 days to deliver. Expected demand during the wait is 8 multiplied by 7, which is 56 candles. The shop keeps 20 candles as safety stock for a late delivery or a busy weekend. Its reorder point is therefore 76 candles. When stock falls to 76, the shop starts the order. It does not order 76 candles; it orders enough to cover the next sales period.
Safety stock should vary by product. Give more buffer to best sellers and unreliable suppliers, and less to slow movers. Ask the AI to calculate a reorder point per product, then adjust safety stock yourself.
Four prompt recipes you can copy and reuse
Good prompts give the AI your real numbers and ask for the working to be shown. Paste your sheet as plain text, remove customer names and personal details first, and tell the AI not to invent any missing figure.
Prompt 1, weekly reorder check: “From this stock table, calculate inventory position (units on hand plus units on order) and weeks of cover for each product. Required cover is lead time plus two safety weeks. Mark each product as order now, watch or fine, and show the maths.”
Prompt 2, slow stock check: “List every product with no sales in the last 60 days, show units on hand and total cost tied up, and suggest whether to discount, bundle or stop stocking each one.”
Prompt 3, seasonal forecast: “Using the last 12 months of weekly sales, estimate demand for the next 8 weeks. Flag products where demand rises before Christmas and Black Friday, and show extra units needed above the normal weekly average.”
Prompt 4, order draft: “For products marked order now, calculate the quantity needed to cover lead time plus safety stock, adjust for minimum order quantity and case packs, and total the cost. Stop if the total is above my cash limit of [amount] and tell me which items to defer.”
Run prompt 1 weekly, prompt 2 monthly, prompt 3 before each seasonal peak, and prompt 4 only after checking the first three. If an output looks wrong, ask the AI to show its source rows.
Plan for UK seasonality, not average demand
Average demand misleads UK retailers because trade is seasonal: quiet January, a lift around Easter, summer peaks for outdoor and holiday goods, a Black Friday surge in late November, the Christmas peak in December, then returns and clearance in January.
Tell the AI the event, the weeks it covers and last year lift. For example: “Last November, gift set sales rose by 60 percent compared with October. Estimate stock for the four weeks before Christmas, allowing for a 10 day lead time and one week of safety stock.”
Also allow for bank holidays and supplier closures. Many small suppliers close between Christmas and New Year. Flag any product whose lead time crosses a closure, and order it a week earlier. A reorder point based on a normal 7 day lead time fails when the supplier is shut for 10 days.
Use the same forecast in your wider planning; see our guide on how to use AI to write a business plan for UK founders.
Clear slow movers and keep AI honest
Slow stock is cash sitting on a shelf. Each month, take the slow list from prompt 2 and choose one action per product: discount it, bundle it, move it, return it if allowed, or stop stocking it.
Before every order, run three checks. Count your top products by value and sales volume, and fix the sheet if it says 40 units but the shelf holds 28. Confirm supplier lead times, minimum orders and prices, since an order built on last quarter prices understates cost. Remove customer names, addresses and emails before pasting any export into a chatbot, and keep final stock valuations in your accounting system. Our roundup of the best AI bookkeeping and accounting tools for small businesses covers options that handle stock value and VAT records for UK firms.
Frequently asked questions
Can a small business use AI for inventory management without buying software?
Yes. If you can export a spreadsheet of products, sales and supplier lead times, a free chatbot can calculate reorder points, flag slow stock and draft order lists. Start with one weekly check and verify the maths.
How much sales history does AI need to forecast demand?
For a basic reorder check, 8 to 13 weeks is enough. For seasonal planning, use a full 12 months so Christmas, Easter and summer peaks are visible. With less history, ask for a conservative forecast and review it weekly.
What is the difference between a reorder point and an order quantity?
The reorder point is the stock level that triggers an order. The order quantity is how much you buy once triggered. Set the point from daily demand, lead time and safety stock, then set the quantity from stock duration, supplier minimums, case sizes and your cash limit.
How do I stop AI from inventing stock figures?
Tell it not to invent missing numbers, ask it to show working row by row, and check output against your sheet before ordering. If a figure has no source row, treat it as missing.
Should I let AI place orders automatically?
No, not for a small business. Let AI draft and check the order, but keep a person approving every purchase. Automatic ordering only makes sense once counts are accurate and the system has passed a seasonal peak without errors.
Conclusion
AI earns its place in a small stockroom by doing repetitive checking owners rarely have time for: sorting sellers, calculating reorder points and drafting order lists from a simple spreadsheet. The advantage comes from a clean sheet, a fixed weekly routine, prompts that show working, and a human check before every order. Start with one product category this week.
