AI is replacing the click with an answer, recommendation and, increasingly, a completed transaction. That threatens retailers’ audiences and Google’s traditional advertising model – but it could also turn commerce media into the mechanism through which machines decide what people buy.
For the past two decades, digital commerce has been organised around a remarkably durable sequence: someone searches, clicks, visits a website and eventually buys something.
Google monetised the first part of that journey. Retailers monetised the second. Retail media networks used the resulting traffic and transaction data to sell advertising closer to the point of purchase.
Three developments now threaten to pull that structure apart.
The first is “Google Zero”: the growing tendency for Google to answer questions without sending users elsewhere. The second is agentic commerce, in which AI does not simply recommend products but compares, selects and potentially purchases them. The third is the European Commission (EC) decision requiring Google to share anonymised search data with rival search engines and qualifying AI chatbots.
Together, they point towards a commerce environment in which considerably fewer consumers visit retailer websites – and in which Google may no longer have exclusive control over the intelligence needed to intercept them.
From zero-click search to zero-visit commerce
Google Zero is already more than a theoretical concern. Research published in June 2026 by SparkToro, based on Similarweb clickstream data suggests 68% of US Google searches ended without a click during the first four months of 2026, up from 60.45% in 2024. AI Overviews appeared in more than 20% of searches and, when present, reduced clickthrough rates by almost 60%. These figures require some methodological caution, but the direction is unmistakable.
Agentic commerce takes that trend much further. An AI answer may remove the need to visit a publisher; an AI shopping agent could remove the need to visit a retailer.
New WARC and PHD research forecasts that agent-facilitated consumer spending will rise from $944bn in 2026 to $3.35trn in 2030, equivalent to 3.8% of global consumer expenditure. The UK alone is expected to account for $131.2bn.
Retail-related agentic spending is forecast to grow 218.7%, from $62.7bn to $199.9bn. Food will reach $292.8bn, soft drinks $304.8bn and alcoholic drinks $198.4bn. These are precisely the frequent, data-rich and replenishment-led categories on which many retail media networks are being built.
The danger is not simply that an AI agent might choose a different brand, it is that the consumer may never enter the retailer’s digital environment at all.
The agent could compare prices, availability, delivery times, product specifications, reviews and loyalty benefits across multiple retailers, then place the order with whichever supplier best satisfies its instructions. The retailer receives the transaction but loses the visit, the browsing signals, the opportunity for discovery and much of the onsite advertising inventory.
As WARC warns, retailers risk becoming fulfilment providers rather than destinations.
Does Google risk destroying its own ad business?
But there a real paradox at the heart of this move. If Google answers the question, selects the product and completes the purchase, consumers will have fewer reasons to click either organic results or conventional search advertisements – including Google’s own.
Zero-click does not necessarily mean zero-advertising revenue. Google’s strategy is to rebuild advertising inside the AI journey. It is already testing sponsored recommendations and “Highlighted Answers” within AI Mode, alongside Direct Offers that surface discounts when a shopper appears ready to buy. Native checkout enables participating retailers to complete purchases without requiring the shopper to leave Google.
Its Universal Commerce Protocol is designed to connect AI surfaces with retailers, payment providers and post-purchase services. Google’s agentic commerce announcement describes a system in which retailers remain seller of record, but Google controls much of the surrounding experience.
Google is therefore not necessarily sacrificing advertising, it is attempting to migrate from selling clicks to selling inclusion, recommendation, offers and completed outcomes. Its newer formats already place sponsored products within AI-generated responses and attach AI-written explanations to them. Google’s 2026 advertising announcements make that direction explicit.
In effect, the search results page is becoming a commerce-media environment. Google supplies the audience, intent signal, recommendation engine, advertising auction and potentially the checkout. That looks less like traditional search advertising and more like a vast, cross-retailer retail media network.
The risk for Google is that it must manage a delicate question of trust: can an AI claim to be acting for the consumer while sponsored interventions influence its recommendations? Paid placement will need to remain distinguishable from the agent’s supposedly independent judgement.
Europe weakens the data advantage
The European Commission’s ruling complicates Google’s attempt to dominate this new layer. Under binding measures issued through the Digital Markets Act, Google must share anonymised search data with eligible competitors on fair terms. Crucially, the Commission has confirmed that AI chatbots offering search functionality can qualify and should receive, subject to anonymisation, the same types of data Google uses to optimise its own search services.
That potentially gives rival AI services access to patterns within ranking, query, click and viewing behaviour collected at a scale no competitor could easily reproduce.
The ruling does not hand rivals Google’s advertising system, identifiable user profiles or complete commercial intelligence. Access will be controlled, anonymised and chargeable. Nevertheless, it reduces one of Google’s structural advantages: the feedback loop created by billions of searches and user responses.
Combined with separate EU measures giving rival AI assistants greater access to Android functions, it could help ChatGPT, Perplexity and others become credible shopping gateways. Instead of commerce consolidating around Google’s agent, Europe could create competition among agents – each trying to become the consumer’s preferred route to products and services.
Retail media must start marketing to machines
For retailers and brands, the strategic priority consequently shifts from winning the click to entering the machine’s shortlist.
That requires accurate, structured and continuously updated product data covering price, availability, delivery, returns, sustainability, compatibility, ingredients and loyalty benefits. Retailers will also need APIs capable of exposing inventory, accepting agent-led orders and returning reliable transaction signals.
Commerce media will follow. Sponsored products will not disappear, but they may be bought across AI recommendation environments rather than only on retailer websites. Measurement will move away from page impressions and clickthrough rates towards shortlist inclusion, recommendation share, incremental selection and completed agent-assisted sales.
First-party transaction data becomes more valuable, because retailers can prove what was ultimately bought. Yet audience ownership becomes weaker if the agent controls discovery and customer interaction.
Brand also matters more than the automation rhetoric suggests. Agents will optimise routine purchases, but they will still operate within consumer preferences. A shopper who instructs an agent to buy “my usual Coke” or “a reliable Bosch dishwasher” has already embedded brand equity into the algorithm.
Retail media is therefore not facing extinction, rather facing a change in customer. Increasingly, the immediate audience for an offer, product listing or sponsored recommendation will not be a person scrolling a page. It will be a machine deciding what that person gets to see – and perhaps what they ultimately buy.
You are in: Home » Retail Media » Google Zero – retail media’s clickless reckoning?
Google Zero – retail media’s clickless reckoning?
Paul Skeldon
AI is replacing the click with an answer, recommendation and, increasingly, a completed transaction. That threatens retailers’ audiences and Google’s traditional advertising model – but it could also turn commerce media into the mechanism through which machines decide what people buy.
For the past two decades, digital commerce has been organised around a remarkably durable sequence: someone searches, clicks, visits a website and eventually buys something.
Google monetised the first part of that journey. Retailers monetised the second. Retail media networks used the resulting traffic and transaction data to sell advertising closer to the point of purchase.
Three developments now threaten to pull that structure apart.
The first is “Google Zero”: the growing tendency for Google to answer questions without sending users elsewhere. The second is agentic commerce, in which AI does not simply recommend products but compares, selects and potentially purchases them. The third is the European Commission (EC) decision requiring Google to share anonymised search data with rival search engines and qualifying AI chatbots.
Together, they point towards a commerce environment in which considerably fewer consumers visit retailer websites – and in which Google may no longer have exclusive control over the intelligence needed to intercept them.
From zero-click search to zero-visit commerce
Google Zero is already more than a theoretical concern. Research published in June 2026 by SparkToro, based on Similarweb clickstream data suggests 68% of US Google searches ended without a click during the first four months of 2026, up from 60.45% in 2024. AI Overviews appeared in more than 20% of searches and, when present, reduced clickthrough rates by almost 60%. These figures require some methodological caution, but the direction is unmistakable.
Agentic commerce takes that trend much further. An AI answer may remove the need to visit a publisher; an AI shopping agent could remove the need to visit a retailer.
New WARC and PHD research forecasts that agent-facilitated consumer spending will rise from $944bn in 2026 to $3.35trn in 2030, equivalent to 3.8% of global consumer expenditure. The UK alone is expected to account for $131.2bn.
Retail-related agentic spending is forecast to grow 218.7%, from $62.7bn to $199.9bn. Food will reach $292.8bn, soft drinks $304.8bn and alcoholic drinks $198.4bn. These are precisely the frequent, data-rich and replenishment-led categories on which many retail media networks are being built.
The danger is not simply that an AI agent might choose a different brand, it is that the consumer may never enter the retailer’s digital environment at all.
The agent could compare prices, availability, delivery times, product specifications, reviews and loyalty benefits across multiple retailers, then place the order with whichever supplier best satisfies its instructions. The retailer receives the transaction but loses the visit, the browsing signals, the opportunity for discovery and much of the onsite advertising inventory.
As WARC warns, retailers risk becoming fulfilment providers rather than destinations.
Does Google risk destroying its own ad business?
But there a real paradox at the heart of this move. If Google answers the question, selects the product and completes the purchase, consumers will have fewer reasons to click either organic results or conventional search advertisements – including Google’s own.
Zero-click does not necessarily mean zero-advertising revenue. Google’s strategy is to rebuild advertising inside the AI journey. It is already testing sponsored recommendations and “Highlighted Answers” within AI Mode, alongside Direct Offers that surface discounts when a shopper appears ready to buy. Native checkout enables participating retailers to complete purchases without requiring the shopper to leave Google.
Its Universal Commerce Protocol is designed to connect AI surfaces with retailers, payment providers and post-purchase services. Google’s agentic commerce announcement describes a system in which retailers remain seller of record, but Google controls much of the surrounding experience.
Google is therefore not necessarily sacrificing advertising, it is attempting to migrate from selling clicks to selling inclusion, recommendation, offers and completed outcomes. Its newer formats already place sponsored products within AI-generated responses and attach AI-written explanations to them. Google’s 2026 advertising announcements make that direction explicit.
In effect, the search results page is becoming a commerce-media environment. Google supplies the audience, intent signal, recommendation engine, advertising auction and potentially the checkout. That looks less like traditional search advertising and more like a vast, cross-retailer retail media network.
The risk for Google is that it must manage a delicate question of trust: can an AI claim to be acting for the consumer while sponsored interventions influence its recommendations? Paid placement will need to remain distinguishable from the agent’s supposedly independent judgement.
Europe weakens the data advantage
The European Commission’s ruling complicates Google’s attempt to dominate this new layer. Under binding measures issued through the Digital Markets Act, Google must share anonymised search data with eligible competitors on fair terms. Crucially, the Commission has confirmed that AI chatbots offering search functionality can qualify and should receive, subject to anonymisation, the same types of data Google uses to optimise its own search services.
That potentially gives rival AI services access to patterns within ranking, query, click and viewing behaviour collected at a scale no competitor could easily reproduce.
The ruling does not hand rivals Google’s advertising system, identifiable user profiles or complete commercial intelligence. Access will be controlled, anonymised and chargeable. Nevertheless, it reduces one of Google’s structural advantages: the feedback loop created by billions of searches and user responses.
Combined with separate EU measures giving rival AI assistants greater access to Android functions, it could help ChatGPT, Perplexity and others become credible shopping gateways. Instead of commerce consolidating around Google’s agent, Europe could create competition among agents – each trying to become the consumer’s preferred route to products and services.
Retail media must start marketing to machines
For retailers and brands, the strategic priority consequently shifts from winning the click to entering the machine’s shortlist.
That requires accurate, structured and continuously updated product data covering price, availability, delivery, returns, sustainability, compatibility, ingredients and loyalty benefits. Retailers will also need APIs capable of exposing inventory, accepting agent-led orders and returning reliable transaction signals.
Commerce media will follow. Sponsored products will not disappear, but they may be bought across AI recommendation environments rather than only on retailer websites. Measurement will move away from page impressions and clickthrough rates towards shortlist inclusion, recommendation share, incremental selection and completed agent-assisted sales.
First-party transaction data becomes more valuable, because retailers can prove what was ultimately bought. Yet audience ownership becomes weaker if the agent controls discovery and customer interaction.
Brand also matters more than the automation rhetoric suggests. Agents will optimise routine purchases, but they will still operate within consumer preferences. A shopper who instructs an agent to buy “my usual Coke” or “a reliable Bosch dishwasher” has already embedded brand equity into the algorithm.
Retail media is therefore not facing extinction, rather facing a change in customer. Increasingly, the immediate audience for an offer, product listing or sponsored recommendation will not be a person scrolling a page. It will be a machine deciding what that person gets to see – and perhaps what they ultimately buy.
Read More
You may also like
Subscribe to our email community