Your supplier is about to start ordering for you. Handy, right up until you want to switch

Metcash, the $19.6 billion wholesaler behind IGA and a long tail of independent retailers and corner stores, is looking at agentic AI that would place its customers’ routine orders for them, reported by iTnews.

The example its General Manager of B2B Digital, Simon Williams, gave is worth picturing. A small takeaway shop logs on every Monday at 10:30 at night and orders more or less the same things it ordered last Monday. Software could simply do that, and the owner gets their Monday night back.

It is exploratory for now, and Metcash is being coy about the detail, citing IP. But the direction is clear enough, and it is a different shape to most AI stories we cover here. This one does not require you to adopt anything. It arrives whether you were interested in AI or not, because your supplier adopted it.

They can see how long you spend ordering

The line from Williams that stopped me was about what Metcash already knows: “We know how long they log onto the platform.” They can see the friction in your ordering process, measured in minutes, per customer.

Think about what that means competitively. Your wholesaler knows exactly how much of a chore ordering from them is, and removing that chore is worth real money to them. Their stated goal is not just customer happiness, it is more loyalty and more order volume against their competitors.

That is not sinister. It is good business, and the convenience is genuine. But it is worth understanding what you are being offered, because the friction they are removing is doing a job you may not have noticed.

The Monday night that quietly disappears

Here is the bit worth sitting with.

That 10:30pm Monday order is annoying. It is also the moment, once a week, where a business owner looks at what they are buying and what it costs. Most of the time nothing comes of it. But that is the moment where you notice a price has crept up, or think you should ring around, or remember the other supplier who was cheaper on dairy.

Automate the order and you do not just remove the chore. You remove the decision point. The order still happens, from the same supplier, at whatever price is current, and the moment where you might have chosen otherwise never arrives.

That does not make automated ordering a bad deal. For a lot of businesses the time saved is worth more than the sharpened pencil, and that is a perfectly rational trade. It only becomes a bad deal if you make it by accident.

So make it deliberately. If your supplier offers to run your ordering, say yes if the maths works, and then put a recurring reminder in your calendar to do quarterly what you were previously doing weekly by accident: check the prices, check the alternatives, make sure the automation is still serving you rather than just serving them.

Now flip it around

The other half of this story is the one that actually affects most of our clients, because most Melbourne businesses are not corner stores buying from Metcash. They are the ones being bought from.

If a wholesaler’s software can place an order without a human involved, then before long software will be placing orders with you. Not necessarily a dramatic AI agent. Just a customer’s system, or their bookkeeper’s system, or an assistant acting for them, trying to complete a purchase or a booking without a person in the loop.

Most SME intake is not built for that at all. Orders arrive by phone call and get written on a pad. Quotes go out as a PDF attached to an email. Reordering means someone remembers to ring. There is nothing on the other end that a machine could use, so the machine goes somewhere that has one.

We made this case about consumer bookings when Airtasker made itself completable inside ChatGPT. Metcash is the same shift arriving through the trade door instead of the front one.

What to actually do about it

Nothing exotic, and nothing that only pays off if the agentic future arrives on schedule. All of this is worth doing anyway:

1. Make repeat ordering possible without you. A standing order, a reorder link, a saved cart, anything that lets a regular customer buy again without a phone call. Your best customers are the ones buying the same thing repeatedly, and they are the easiest to lose to whoever makes it easier.
2. Publish what things cost, even roughly. A price a machine can read is a price it can act on. “Call for a quote” is invisible to software and increasingly to people.
3. Give your intake a front door that is not an inbox. A real form or booking path that lands somewhere structured, so an order arriving at 10:30 on a Monday night gets captured properly instead of sitting unread until Wednesday.
4. Know what your own suppliers can see. If you are on someone’s platform, assume they have the same data Metcash has, and price accordingly.

This is the same unglamorous plumbing we argued the parliamentary inquiry into SME AI adoption should be funding. It is not exciting. It is what makes everything else possible.

The func.digital take

The interesting thing about the Metcash story is that nobody in it is doing anything wrong. A wholesaler is removing friction for its customers, which is what good suppliers do. Retailers get their Monday nights back, which they should welcome.

It is just worth going in with your eyes open on both counts. Take the convenience, and keep the decision. Then look hard at your own side of the counter, because the same logic that lets your supplier order for you will shortly let your customers’ systems order from you, and the businesses that are easy to buy from automatically will quietly take share from the ones that are not.

If you want to know whether a machine could place an order with you today, or whether everything still routes through someone’s memory and an inbox, that is one of the things we look at in a free digital systems audit. Get in touch, and let’s make your business easy to buy from.