Too much on the shelf is dead money. Too little is a “no” to a customer who had already walked in.
Purchasing and stock is a balance that never settles. On one side, money tied up in boxes nobody asks for; on the other, the customer who walks away because the item will be “back in tomorrow”. Most small and medium-sized businesses (SMEs) run this on gut feeling and on the experience of one person who knows how things stand. That works, until that person is on holiday or leaves. It is exactly the kind of decision where a system that looks at the numbers every day makes the difference, with a buyer who makes the final call.
// A report from a fictional technical wholesaler with 19 employees, 6,000 items in stock and one buyer who keeps everything “in his head”. The company is fictional, the situations are real: we run into them again and again.
Sound familiar?
- Reordering is done on gut feeling: the buyer walks past the shelves and orders whatever looks “nearly out”.
- The bestseller has just sold out; the customer orders it online somewhere else and does not come back.
- In the warehouse sits a pallet that was already there last year. Nobody dares to write it off.
- Every supplier has its own ordering portal, its own minimum order quantities and its own lead time in someone’s head.
- The online shop says “in stock” while the last one in the shop has just been sold.
- Supplier price increases arrive as a PDF and are only entered into the system weeks later.
- Once a year the big stocktake, with half the company in on a Saturday and a result nobody trusts.
- The person who usually places the orders is off sick, and nobody knows exactly what has to go out the door this week.
What usually runs here
The software we come across here most often. It does not have to go; we build around it and in between.
A day in purchasing & stock: now and next
On the left, the day as it usually runs now. On the right, the same day once the dull parts are automated and AI does the groundwork. Watch what changes: not the people, but where their time goes.
- 07:45BuyerWalks through the warehouse with a cup of coffee and jots down on a slip of paper what is nearly out. What sits at the back, he does not see.
- 09:30SalesCustomer at the counter for two boxes of couplings. “Sold out, back in on Thursday.” The customer walks to the shop down the road.
- 11:00BuyerWorks through four supplier portals, each with its own login and minimum order value. Orders a bit extra, just to be safe.
- 13:15WarehouseFinds three boxes behind a rack that have been there for a year. The item has been dropped from the range. They stay put, because writing them off feels like throwing money away.
- 14:40Online shopA customer calls: ordered online, but it turns out not to be in stock after all. The shop and the online shop are not counting the same stock.
- 16:20Managing directorAsks how much money is tied up in stock. Answer: “A lot.” There is no exact figure, and for half of it nobody knows whether it still sells or not.
- 17:30BuyerPuts a supplier’s price list PDF on the “enter next week” pile. Until then, the margin on those items is wrong.
- 07:45BuyerOpens last night’s order proposal: per item the expected demand, the supplier’s lead time and what is already on its way. Reordering means giving approval, not searching.
- 09:30SalesThe couplings are on the shelf: the system had seen demand picking up and had already reordered. The customer pays and comes back.
- 11:00BuyerAll orders go out through one screen, bundled per supplier to get above the minimum order value. Lead times come from the history, not from memory.
- 13:15WarehouseThe three idle boxes are on the “get it moving” list: with a proposal for a promotion or a return to the supplier. Dead money becomes visible and gets dealt with.
- 14:40Online shopShop and online shop count the same stock. If the shop sells the last one, the online shop immediately switches to “available to order with lead time”, not “in stock”.
- 16:20Managing directorSees on his screen: the value of the stock, which part moves fast and which part sits still, and how many “lost sales” there were this week.
- 17:30BuyerThe supplier’s price list has been read in as a file; the new purchase prices and margins are ready with “approve?”. No more pile.
Where AI makes the difference here
Per use case: what the automation does, what a person keeps doing, and whether it is proven or still promising. We say which one honestly.
Order proposals based on demand
The system forecasts demand per item from your own sales history, season and lead time, and prepares an order proposal: how much, from whom, when.
Still human: the buyer approves, adjusts for what he knows (an order that is coming in, a supplier that is faltering) and makes the call.
Making lost sales visible
Every time a customer asks for something that is not there, it gets recorded: which item, how often, in which shop or channel. You manage by what you miss, not only by what you sell.
Still human: decides what should structurally be in stock and what you deliberately keep on order.
Tracking down dead money
Items that have been sitting still for a long time or are becoming obsolete automatically surface with a proposal: promotion, bundle, mark down or return to the supplier.
Still human: chooses the approach; the AI works out what each choice yields in space and money.
Stock that is right everywhere
Shop, online shop, warehouse and bookkeeping count the same stock. Sell on one channel and the rest moves with it; the online shop never again promises what is not there.
Still human: resolves the exceptions (breakage, returns, count differences) that the system flags.
Supplier lists without retyping
Price lists and order confirmations (PDF, Excel or e-order via Peppol) are read in: new prices, lead times and item numbers arrive as a proposal.
Still human: approves the new prices and margins before they go live.
Stocktaking without the Saturday shift
Instead of one big annual stocktake, the system has small counts done continuously on the items that matter (cycle counting), with scanners.
Still human: works through that day’s count list; half the company no longer has to come in on a Saturday.
Honestly: what AI does not do here
- A forecast is not a crystal ball. With a new item without history, or with a sudden spike, the system gets it wrong too. That is why it proposes and the buyer decides, especially at the start.
- Automatic ordering only works if your basics are right: one item is one item, not three variants with a typo. First clean up your item master, then automate. We do that clean-up together.
- Less stock is not always the goal. Sometimes deliberately holding a bit more is actually smarter, because a lost sale costs you more than the storage. The system makes that trade-off visible; you make the choice.
- The knowledge of your regular buyer is worth its weight in gold. This does not replace him, it captures his gut feeling in numbers, so the company does not grind to a halt when he is away for a while.
The objections we hear — and whether we can close them off
We looked them up and heard them from clients. For each objection we say whether it is really solvable, partly solvable, or a risk that stays and that you accept knowingly.
Can it be closed off: yes. Great, and fragile. If he is off sick, goes on holiday or leaves, that knowledge walks out the door with him. We capture his way of working in the system, so the company keeps running. He stays in charge; the system is his memory.
- 01Week 1: core session in purchasing and in the warehouse. We follow one ordering round and one customer request that could not be met, and count the steps.
- 02Week 1: switch on lost-sales recording. Costs next to nothing and shows straight away what you are missing out on.
- 03Weeks 2 and 3: order proposals for your best-selling items, with approval by the buyer. Look first, then trust.
- 04Weeks 3 and 4: align stock between shop, online shop and bookkeeping; dead money in view with a clear-out proposal.
- 05Week 4: measure. Lost sales, stock value, number of rush orders. Only then the next step.
First the free AI scan, then a core session on your floor (a half-day of 4 hours, €596 excl. VAT and travel costs) with a core report and an honest go/no-go. We build on a project basis with fixed hours: €110 per hour for straightforward work, €165 for complex development.
More hours only with your written approval. If we cannot deliver, you pay nothing for what was not delivered.
Want to know what this means for your purchasing & stock?
Take the free AI scan or book an intro call with someone who builds it themselves. A reply within one working day, no slides, no obligations.
Most businesses start with the tool and then get stuck. It is better to start with one piece of work that comes back every week and nobody enjoys.
The build costs are usually the easiest part of the bill. The surprises are in the clean-up beforehand, the monthly usage and the maintenance afterwards.