The Requisition That Names No Supplier
When stock runs low our system can raise a purchase requisition on its own. It raises it against a department. It does not name a supplier, look at who supplied the item last time, or consider who is cheapest or quickest — because nothing in the automation has any concept of who you buy from.
The verdict
Our replenishment automation is real, the approval behind it is properly enforced, and it will genuinely stop you running out of things because nobody read a report. What it will not do is help you buy well, because it hands a buyer a requisition with no history on it and leaves them to reconstruct the sourcing picture by hand — which, in a market where sourcing is the actual skill, means the automation has taken the easy half. Ask us for the history on the requisition rather than for the rule that creates it. The rule is the demo; the history is the work.
Automating replenishment is usually sold as removing a decision. It is worth being precise about which decision, because a reorder involves at least three and they are not equally hard.
- When to buy — has stock crossed the line where you need more. This is a comparison, and a computer should obviously do it.
- How much to buy — a policy question, argued separately in the calculator is not the feature. Ours does a subtraction.
- Who to buy it from — the one that requires judgement, carries the price, and determines whether the goods arrive when you were told they would.
Ours automates the first, approximates the second, and does not touch the third at all.
Where the automation stops
The rule fires when an item crosses its reorder point and creates a purchase requisition. That requisition is raised against a department — the rule requires one, and refuses to do anything if it has not been told which. The requisition then goes through approval, which is properly enforced: no purchase order can be raised from a requisition that has not been approved.
So the trail is clean and the control is real. What the requisition does not carry is any indication of who should supply the goods. Not a suggested vendor, not the one who supplied it last time, not the one with the best price on file. The automation has produced a documented request for something, addressed to nobody.
What the machine does, and where the work resumes
Automatic
Runs without anybody watching
- Notices stock has crossed the reorder point
- Works out a quantity by subtraction, or uses a fixed one
- Raises a requisition against a named department
- Puts it into the approval queue
- Refuses to let an order be raised until it is approved
Still manual
A person, every time
- Deciding who to ask
- Checking whether that supplier is still prequalified
- Requesting quotations if more than one is in play
- Comparing what comes back
- Raising the actual purchase order against the chosen supplier
Read the two columns by effort rather than by count. The left column is a comparison and a database write. The right column contains every part of the process where money and reliability are decided.
The automation removes the noticing and leaves the choosing. Noticing was never the expensive part.
Why the sourcing half is the whole job in this region
The value of automating supplier selection depends entirely on how much variation there is between suppliers. Where a part has one credible source, choosing is not a decision and the automation has lost nothing by skipping it.
That is not the situation for most businesses buying in East Asia. Components and finished goods here typically have many credible sources at genuinely different price, lead time and quality positions, and the whole practice of buying well consists of holding that landscape in mind and moving between suppliers as it shifts. Sourcing is not an administrative step before the order; it is the skill.
Which cuts both ways, and honesty requires the second half. In a market where choosing well is that valuable, you would not want the system choosing for you. A rule that automatically ordered from last time's supplier would be worse than one that names nobody, because it would quietly entrench whoever happened to be used first and make the entrenchment invisible.
So the criticism is not that we fail to choose. It is that we do not inform the choice — the requisition arrives with no history attached, and everything the buyer needs to decide well is somewhere else in the system, unassembled.
What the requisition could carry and does not
- Who supplied this item last time, and what was paid.
- Every supplier who has ever quoted for it, and the spread between them.
- How long each of them actually took against what they promised — which our supplier performance record was designed to hold and has never been populated.
- Whether any of them is prequalified today, or has been blacklisted since you last bought.
- Whether there is an agreement in force with an agreed price.
Every item on that list except the third is already in the system somewhere. The buyer assembles it by hand, per requisition, from memory and from screens they have to go and find — which is why in practice most people skip it and buy from whoever they bought from last time. The absence of the information does not remove the decision; it just makes the lazy answer the default.
Four questions about replenishment automation
When the rule fires, who is the requisition addressed to?
What a straight answer sounds like
A supplier, a department, or nobody. Ours is a department.
Why it matters
Automation that stops before sourcing has automated the cheap half and left the expensive half untouched.
Does the buyer see purchase history on the requisition?
What a straight answer sounds like
A screen showing last price and last supplier.
Why it matters
Without it the default is whoever was used last, chosen by memory rather than by comparison.
Would the system stop me ordering from a blacklisted supplier?
What a straight answer sounds like
A block at order entry.
Why it matters
Eligibility recorded but not enforced at the moment of buying is a list rather than a control.
Can a contracted price be applied automatically?
What a straight answer sounds like
Yes with an agreement record, or no.
Why it matters
Negotiating a rate that nobody applies is a common and expensive way to waste a negotiation.
Two, and neither of them chooses for you
We would deliberately stop short of automatic selection, for the reason given above.
Purchase history on the requisition
Last supplier, last price, last actual lead time, and everyone who has quoted for this item before — shown at the moment somebody is deciding rather than filed where they would have to go and look. This is assembly of data we already hold, and it is the whole of the value.
Supply agreements with contracted prices
A recorded agreement with a supplier for an item at a price for a period, applied when an order is raised. This is the case where automatic selection genuinely is correct, because you already made the choice deliberately and wrote it down.
The distinction we would hold to: automate the decision you have already made and recorded, never the decision you have not.
Talk to us about sourcingWhat AWRA OpsHub does today
- A reorder point per item, and a rule that fires when stock crosses it.
- Automatic requisition creation against a department, without anybody watching a report.
- Approval enforced before an order can be raised, so automation feeds a controlled process.
- Requests for quotation and quotation comparison, once a human is driving.
- Prequalification and blacklisting, so eligibility is decided and recorded.
What it does not do
- No supplier on the requisition. None suggested, none defaulted, none carried over from last time.
- No purchase history on the requisition, so who supplied it before and at what price is not in front of the person deciding.
- No supplier performance to consult, because the record that would hold it is never populated.
- No check that a suggested supplier is still prequalified, since there is no suggestion to check.
- No supply agreements or contracted prices, so an agreed rate cannot be applied automatically.
Not ours, by choice
- We make no claims about the supplier landscape, industrial policy or sourcing conditions of any country. That buying well here means moving between many credible sources is a description our customers give us, not a statistic we are publishing.
- We would not build automatic supplier selection as a silent default, and we would argue against it if asked. Entrenching last time's supplier invisibly is a worse outcome than naming nobody.
Tell us how you choose between suppliers
What you look at, and where that information lives today. We will tell you what we can put in front of your buyers now, what would need building, and where we would push back on automating a judgement.
Talk to us about procurement