Agentic commerce, the idea that an artificial intelligence agent can browse, compare, negotiate and complete a purchase on your behalf, has become one of the most hyped concepts in retail. The pitch is seductive: tell your assistant you need running shoes in a certain size and budget, and it does the legwork, checks the reviews, applies your loyalty points and checks out, all without you lifting a finger. Yet a persistent question is now hanging over the whole proposition. Will people actually trust a machine to spend their money and hold their data?
Fresh research reported by Inside Retail suggests the enthusiasm is real but conditional. Consumers are genuinely curious about AI-powered shopping, and many like the promise of saving time and cutting through the noise of endless product choices. What they are far less comfortable with is the trade that makes it work. Handing an agent the keys means handing it payment details, purchase history, personal preferences and, crucially, the authority to act. The findings point to a widening gap between how willing people are to experiment and how willing they are to let go of control.
Why the trust gap matters
The mechanics of agentic commerce are what make it both powerful and unsettling. For an agent to be genuinely useful, it needs context. It has to know your sizes, your budget, your past orders and your tastes, and it often needs stored payment credentials to complete a transaction. That is a far deeper level of access than a conventional shopping app, which largely waits for you to tap the buttons. An autonomous agent is, by design, meant to make decisions without checking in at every step, and that autonomy is precisely what gives shoppers pause.
The research frames this as a control problem as much as a privacy problem. People are not only worried about where their data goes, they are worried about being cut out of the loop. The fear of an agent buying the wrong thing, overspending, locking in a subscription or acting on a misunderstood instruction looms large. It is one thing to let an algorithm recommend a product. It is another to let it reach into your wallet and commit you to the sale.
Two ways to read the findings
The optimistic reading, favoured by many in the technology and retail sector, is that this is a normal adoption curve. New conveniences almost always arrive wrapped in suspicion, from online banking to tap-and-go payments to ride-sharing, and the wariness tends to fade once the benefits become obvious and the safeguards become visible. On this view, privacy concern is not a brick wall but a design brief. Build agents that ask before they buy, cap spending, keep a clear audit trail and let shoppers reverse decisions easily, and the trust will follow. The companies pushing hardest into the space, including the big payment networks and AI platforms rolling out agentic checkout tools, are betting on exactly that trajectory.
The more cautious reading is that agentic commerce asks for something qualitatively different from past conveniences, and that the usual adoption optimism underestimates it. Delegating a purchase is not the same as delegating a recommendation. When an agent acts autonomously, accountability becomes murky. If the wrong item arrives, or a price turns out to be inflated, or data is shared with a merchant the shopper never chose, who is responsible? Consumer advocates argue that until those questions have clear answers, a large slice of the market will keep agents on a short leash, using them to research and shortlist rather than to transact. On that view, the privacy gap is not a passing phase but a structural ceiling on how far the technology can scale.
What it means for Australia
For Australian retailers and shoppers, the timing is pointed. The country is in the middle of its most significant privacy overhaul in decades, with reforms to the Privacy Act progressively tightening the rules on how personal information is collected, used and disclosed. Agentic commerce lands squarely in the path of those changes. An agent that hoovers up purchase history and shares it across merchants to strike a deal is exactly the kind of data flow regulators and the Office of the Australian Information Commissioner are watching closely.
Australian consumers have also grown noticeably more guarded after a run of high-profile data breaches across telecommunications, health and financial services. Confidence that any given company will keep personal data safe is not high, and that scepticism will travel with shoppers when they weigh up whether to let an agent act for them. Retailers hoping to roll out agentic features here cannot simply import an overseas playbook. They will need to demonstrate local compliance, clear consent and genuine user control, or risk the feature being seen as another data grab dressed up as convenience.
There is a commercial dimension too. Australia’s retail sector is fiercely competitive and heavily concentrated, and the big players have spent years building loyalty programs and first-party data assets. Agentic commerce could either supercharge those assets or undermine them, depending on whether the agent works for the shopper or for the merchant. If an independent agent can objectively compare prices across retailers, it threatens the carefully managed customer relationships that chains rely on. That tension will shape how quickly, and how openly, Australian retailers embrace the technology.
What is next
The near-term path is likely to be incremental rather than revolutionary. Expect to see agents deployed first in low-stakes, high-frequency categories such as groceries, repeat consumables and reordering, where a mistake is cheap and easily reversed, before they move into higher-value purchases. Spending limits, confirmation prompts and transparent logs will be the features that do the heavy lifting on trust, and the retailers that lead on those safeguards may well win the early adopters.
The broader test is whether the industry can close the gap the research identifies, between people who like the idea of AI shopping and people who are willing to actually hand over control. Convenience alone has rarely been enough to overcome a deficit of trust. For agentic commerce to move from novelty to norm in Australia, shoppers will need to believe that the agent is genuinely on their side, that their data is protected, and that they can step in whenever they want. Until then, the technology’s biggest obstacle is not capability. It is confidence.
Sources: Inside Retail.
