Who Will Own the Customer When AI Does the Shopping?
Almost half of all Australians used generative AI in the past year, up from just over a third in 2023. They are using it to research products, compare options, summarise reviews and increasingly to make purchase decisions. Google and Ipsos found that 55% of Australian consumers are excited about AI-powered personal assistants that can help them shop online, plan trips and manage their schedules. Australian consumers are becoming more sophisticated in how they find, evaluate and choose products.
The retailers selling to them, in many cases, are not keeping pace. Last year, the Department of Industry concluded that only approximately one-third of SMEs have adopted AI in any meaningful way. Deloitte reported that two-thirds of Australian SMEs use AI in some form, but only 5% derive substantial value from it. This is a familiar pattern for anyone who has followed the broader AI maturity story in retail: high experimentation, low impact. Lots of pilots, very little production-grade capability.
That gap between consumer sophistication and retailer readiness is about to be increasingly tested and without action from retailers, widened. This article looks at how agentic commerce is emerging, where power may shift in the retail value chain and what Australian retailers should be considering now.
The Rise of the AI Shopper
Agentic commerce is increasingly entering the retail vocabulary. It describes a model of online shopping where an AI agent acts on behalf of a consumer, searching across retailers, comparing products, evaluating pricing and completing purchases with limited or no manual input. Rather than a human browsing a website, scrolling through categories and adding items to a cart, an AI agent receives a brief (“find me running shoes under $200 with good reviews for flat feet”) and executes the entire process autonomously.
The scale of the opportunity is becoming clearer. Morgan Stanley estimates that agentic shoppers could capture 10% to 20% of US e-commerce market share by 2030, representing up to US$385 billion in spending. Bain reports that for some retailers, AI already drives up to 25% of referral traffic, though it remains less than 1% of total traffic today. According to Adobe, one in six purchases during the 2025 Black Friday period were AI-assisted. The current base is small, but the growth trajectory and the investment behind it are significant.
The infrastructure is being assembled by the largest technology and payments companies in the world, right now. Mastercard launched its Agent Pay technology in 2025, enabling verified AI agents to make transactions on behalf of consumers and businesses. Google launched its Universal Commerce Protocol in January 2026, a new open standard for agentic commerce that works across the entire shopping journey, co-developed with Shopify and endorsed by Mastercard, Visa, Walmart, Target and more than twenty other major companies.
Where the Power Shifts
The implications go beyond a new sales channel. Agentic commerce will start to change who controls the customer relationship and with it, challenge the balance of power in the retail value chain.
Retailers have built their competitive position on owning the customer touchpoint. Stores, websites, apps, loyalty programmes, direct marketing, are all mechanisms for maintaining a direct relationship with the person making a buying decision. Agentic commerce introduces a new intermediary. The AI agent does not browse, does not respond to visual merchandising and does not see banner ads. It evaluates data: product descriptions, specifications, pricing, availability, delivery times, return policies, ratings and reviews. When a consumer delegates purchasing to an agent, the retailer’s relationship shifts from the person to the machine.
This pattern has played out before. When a coordination layer sits between a business and its customer, the business controlling that layer captures a disproportionate share of the value. Google did it with search. Agentic commerce is creating new gatekeepers: AI platforms like OpenAI and Google controlling discovery and conversation, payments networks like Mastercard and Visa providing the trust and transaction infrastructure agents need to operate. Retailers lacking visibility in this layer risk becoming fulfilment utilities, selected on price and availability rather than brand or relationship.
Research from Mirakl found that 42% of customers already abandon purchases due to insufficient product information. In an agent-mediated world, the agent is more likely to silently exclude that retailer from its recommendation set. Industry research suggests that where product data is clean and well-organised, AI-generated recommendations can achieve significantly higher conversion rates than traditional search. Data quality is already a differentiator. Agentic commerce turns it into a condition of entry.
Traditional customer relationship assets do not lose all value. Strong loyalty programmes, engaging in-store experiences and first-party customer data remain powerful. Leading retailers have spent years building these assets: using loyalty data to understand preferences and behaviour, investing in physical experiences that create differentiation and developing direct relationships that generate rich, first-party insight. Bain’s research found that consumers currently trust retailers’ own AI agents three times more than third-party agents. There is a window for retailers to leverage these strengths, but it depends on also being visible and competitive in the agent-mediated channel.
The Readiness Question
For Australian retailers, the practical challenge is immediate, but not new. Agentic commerce does not introduce a new problem. It raises the stakes on an existing one. Many retailers are still working through foundational data issues: fragmented customer data, inconsistent product taxonomies, siloed inventory systems and limited API infrastructure. These are not new problems. They held back personalisation, omnichannel fulfilment and operational AI for years.
Every investment a retailer has made in clean product data, well-organised catalogues, reliable fulfilment and accessible systems becomes more valuable in an agent-mediated world.
What is emerging in industry discussions is the concept of “agent engine optimisation”: the discipline of making products discoverable, comparable and trustworthy to autonomous shopping agents. If a customer’s agent is instructed to prioritise retailers with transparent return policies, verified stock levels and competitive delivery times, retailers without machine-readable data on those dimensions will never be evaluated, regardless of how well their website performs for human visitors. The retailers who have been doing this foundational work, often quietly, are the ones best positioned. Agentic commerce validates the strategy they already have.
For those who have not made these investments, the window is narrowing. Consumer AI adoption in Australia is accelerating. The infrastructure for agent-driven commerce is being built. And the shift from human-browsed to agent-evaluated retail will not announce itself. It will show up gradually and quietly.
The question for retail leaders is direct. If an AI agent evaluated your product data, your pricing transparency, your fulfilment reliability and your catalogue quality today, would it choose you?
If the answer is uncertain, the time to act is not when agentic commerce arrives at scale. It is now.
Sources
Google and Ipsos, “AI Adoption in Australia,” January 2025
Department of Industry, Science and Resources, AI Adoption Analysis, June 2025
Deloitte, Australian SME AI Adoption Report, late 2025
Morgan Stanley Research, Agentic Commerce Market Estimates, 2025
Bain & Company, “Agentic AI in Retail: How Autonomous Shopping Is Redefining the Customer Journey,” November 2025
Adobe, AI-Assisted Black Friday and Cyber Week Purchasing Data, November 2025
Mastercard, Agent Pay and Agentic Token Framework, 2025
Google and Shopify, Universal Commerce Protocol, January 2026
McKinsey & Company, “The Agentic Commerce Opportunity,” October 2025
Mirakl, Product Data and Purchase Abandonment Research, 2025