From ten shots at the sale, to just one.
Part one of a multi-part series on agentic commerce for ANZ retailers.
In May, the RX Group team described the agentic commerce trap and asked "who wins"?
But, before a champion can be crowned, there are four tasks every retailer has quietly owned by default for decades that need untangling: being found, being known, getting paid, and being accountable for what happens in between.
This article explores the first of those two tasks, and we'll save the others for next time.
Task One: Being Found
Last October, OpenAI and Walmart put one of the world's largest stores inside a chat window. No app. No search bar. No results page. You typed a keyword for a product that you wanted, and it went and found it for you.
For thirty years, digital commerce ran on one basic loop: a customer had a need, typed it into a search bar, scrolled past the ads, compared a few options, and clicked buy. Our entire craft as merchants was built around that loop. SEO, retargeting, abandoned cart flows, a product page persuasive enough to hold someone's attention. We were competing for the click.
That loop assumed a human was doing the clicking, which is the assumption that is now breaking as Agentic commerce becomes more prominent.
Ask an agent to find you hiking boots and it doesn't hand back ten options with sponsored listings on top. It hands back one, maybe two recommendations. It has already formed a view on terrain, climate, your size, what you paid last time and which retailers are able to deliver to you, before it even asks you a single question.
If your products aren’t included in that shortlist, you're outright invisible - there's no page two and no analytics to tell you how close you are to the top.
Task Two: Being Known
Loyalty programmes, points cards, email open rates: that machinery was built for a world where the customer did the remembering. The agent doesn't need it, it already knows the customer - but does it know you, the retailer?
The agent can't see the stock in your Auckland store, the fit note your best salesperson gives every second customer, the fact that this jacket runs small, or whether you really can deliver by Friday. This is the business-specific information that's sitting in someone's head, on an ageing blog post or a spreadsheet nobody owns.
This means that structured, accurate, machine-readable product knowledge is no longer just housekeeping, it's absolutely critical. It now decides whether you'll even been seen.
What Can You Do About It?
The retailers getting ahead aren't doing anything mystical. They're simply treating "does an agent know who we are" as seriously as they once treated "do we rank on page one."
Here's a test you can run this week. Pick a product - put your customer hat on and ask an agent to find it. You’ll then compare the details an agent returns on that product with your admin backend to ensure:
the specifications displayed are correctly
the stock is flagged as available for a certain size and the store is accurate
your delivery promises can be achieved.
Do it for your top twenty products and see how your best-selling items are displaying today. Then, repeat it as a customer for your three closest competitors to see how you stack up against them.
What comes back is your real benchmarked position, not your assumed one. It's usually worse than you expect, and the gap between you and whoever scores better is down to data quality.
It doesn't stop here though. There are three more stages you can implement to set your organisation up for success:
Reconcile your systems: ensure you have one source of truth per data point. An agent won't forgive an inconsistency the way a human shopper will.
Build the structured layer: once the data is reconciled, expose it so agents can transact against it, not merely discover it. For those of you on Shopify, you’ve got an advantage here as the platform has already done the heavy lifting, shipping a Storefront MCP server and Universal Commerce Protocol support to every store by default.
Govern it: Assign ownership and implement processes that maintain it. This isn't a project you finish. It's an operating model.
The Wrap-Up
Which brings us back to RX Group's original question: who wins? Right now, that's still an open contest. But the retailers likely to win it aren't the ones with the flashiest agent strategy. They're the ones who quietly fixed being found and being remembered before anyone else thought to ask. That shift in seriousness will, more than any single tool, separate the retailers that agents recommend from the ones they quietly route around.
Getting paid and being accountable are the next two tasks on the list, and we'll cover both in a future instalment.
Jo Goldstein, Dan Seddon and Paul Harvey work at Fusion5, a commerce transformation partner working with Shopify retailers across Australia and New Zealand.
Comments