Consumers will take an AI's advice but refuse to let it spend their money, so the real 2026 contest is being the brand an AI trusts enough to recommend. Filip Ivankovic argues Australian operators are over-funding checkout automation shoppers have rejected and under-funding the machine-legible brand authority that decides whether they get named at all.
They will take the advice. They will not hand over the card. A plan that ignores that line is funding the wrong half of the journey.
The Take: Australian operators are pouring budget into automating a checkout that shoppers have already refused, while starving the one thing that decides whether an AI names them at all. The moment of choice has a new gatekeeper in front of it, and it is not a sharper discount or a faster cart. It is whether the machine rates you enough to say your name out loud.
What are consumers actually refusing?
Shoppers want AI to help them choose, not to choose for them. Only 11% will let an AI make the purchase for low-stakes items, so the transaction stays human while the recommendation gets outsourced.
That refusal is the whole story, and most plans are built as if it is not there. The buy-button gets the engineering. Meta showed an assistant at Connect that spends for you. ChatGPT wired in agentic checkout. Retailers are laying agentic commerce rails. Shoppers watched all of it and kept a thumb on the card. Only 11% will let AI make the call even on low-stakes buys like shampoo and bin liners, according to Gartner. The advice they will take. The spending they will not delegate.
That distinction is not a phase to wait out. It is a stable preference, and it splits the year's work cleanly in two. One half automates a step the customer wants to keep. The other half earns the recommendation the customer is happy to accept. Most Australian budgets are funding the first.
share of consumers willing to let AI make even low-stakes purchase decisions, per Gartner
Australia's advertising market will reach A$31.1 billion in 2026, WPP Media forecasts. Retail media is the quickest-growing slice within it at A$2.3 billion, on the same forecast. That is the most bottom-of-funnel money there is, aimed at shoppers already standing at the digital shelf.
Across the Australian businesses we score at New Rebellion, the pattern is consistent: strong Conversion Efficiency and Data & Tracking, a persistent shortfall in Brand & Positioning and the structured, machine-legible credibility that makes a business the one an AI recommends. The gap widens as the automation improves, because a faster cart does nothing for a brand the recommender has never heard of.
For a decade the customer passed you the ball at the moment of choice. You could position yourself to receive it, then close. That lane now has an interceptor standing in it, reading the field and deciding whose name gets released.
Why won't shoppers hand the card to the machine?
Because they have used it and been let down. Accuracy is the first crack. Enough shoppers have caught the AI being wrong that verifying its answers has become a habit, which defeats the point of asking. Then there is the money question. 75% of shoppers say they would lose trust in AI recommendations the moment those results turn out to be paid placements, reports Morningstar. The instant an answer looks bought, the advice is worth nothing to them. That is the trap platforms walk into as they bolt advertising onto the recommender.
Meanwhile 72% say generative AI turns up in their apps and searches whether they asked for it or not, Gartner found. The intermediary is not something a customer opts into. It is the default reader of the field, and it is forming a view on your brand with or without your input.
If the machine does not rate you, you are not in the running. No amount of conversion tactics fixes being invisible to the thing doing the recommending.
What I would do about it
You do not beat an interceptor by sprinting harder at the catch. You beat it upstream, by being the option it trusts enough to release the ball to. For an Australian operator, that means five things.
Make your credibility machine-legible. Structured data, clean entity information and consistent claims across the sources models read matter more than another landing-page tweak. Feed the recommender facts it can parse.
Reinvest from pure conversion into Brand & Positioning. If shoppers do not recognise your name before the query, the AI has no reason to surface it.
Get cited where the models look. Earned coverage, category authority content, verifiable reviews and the comparison sources the models read are what build a recommendable reputation.
Audit your own category. Test it yourself. Ask the major assistants to recommend in your space and watch whether your name shows up at all.
Stop paying for placements that corrode the trust you need. If a sponsored slot reads as bought, it can cost you the credibility that made the recommendation valuable.
How we score, and what these numbers mean
New Rebellion benchmarks Australian businesses on a Marketing Score across dimensions including Brand & Positioning, Conversion Efficiency, Digital Maturity and Data & Tracking. The pattern described here is directional, read from that scoring rather than any single figure. The market data is current to WPP Media's mid-year 2026 forecast and to the Gartner surveys. If you want the full methodology, read how we score.
The brand the machine trusts
Here is my view, put plainly. 2026 is not won at the checkout. It is won earlier, in the read, where the intermediary decides whose name to release to the customer. The operators who spend the year shaving seconds off a cart the AI now controls will wonder why their volume kept sliding. The ones who spend it becoming the answer the machine trusts will still be in the running when the pass comes. Australia's operators have the budget to do both. The question is order of operations, and right now the order is backwards. Win the read, not the sprint. See the benchmarks.
Frequently asked questions
How do I know if an AI recommends my brand?
Test it directly. Ask the major assistants to recommend in your category, note whether your name appears, then track that position over time the way you would track a ranking.
How is this different from SEO?
SEO earns a spot on a page a person scans. This is about being the entity a model trusts enough to name in a single answer, which leans on structured data and third-party credibility rather than keywords alone.
How much should I shift from conversion to brand?
There is no fixed ratio. If your Brand & Positioning lags your Conversion Efficiency by a wide margin, move spend until the recommender knows your name before the query, then hold.
How fast is this changing?
Quickly. With 72% of shoppers already meeting generative AI unprompted, the recommender is the default surface now, not a problem for a future budget.