Google I/O produced a week of coverage on how AI will change the search experience. Most of it has focused on consumer features, with less attention paid to an emerging model for businesses.
Several of the most publicized public demonstrations to I/O take the user from search to action, with Google doing more of the journey in between. While this was a recurring theme throughout I/O, the infrastructure behind these demos was deployed months before the keynote.
Last week’s in-depth analysis argued that the real I/O risk was more economic than technical. This article examines where this economic risk is concentrated and why the business playbook has not caught up with the consumer experience presented by Google on stage.
What Google showed
Google I/O demos included Universal trolley, reservation agent for local services and information agents who monitor announcements or products in the background.
Universal Cart lets you add products into a single cart that persists across all Google surfaces. The booking agent collates prices and availability and provides links to finalize the booking, bringing the trip closer to completion.
However, not all demos were commercial. Google also demonstrated coding, dashboards, simulations and search tools.
The infrastructure was already in motion
I/O made the infrastructure visible to consumers, but it had been in development for a long time.
At the end of 2025, Google deployed payment agentwhich allows Google’s AI to add items to a merchant’s cart and complete purchases.
This year, Google launched the Universal Trade Protocolan open standard for agent commerce. UCP provides agents and merchant systems with a common language, rather than requiring unique logins for each agent.
In April, Sundar Pichai told Stripe CEO Patrick Collison that search would become a “agent manager”. SEJ was follow this change thanks to Google’s agentic search patents and task-based search capabilities since the beginning of the year.
Jay Jaffin, marketing director and strategic advisor at Visor Strategic Advisors, summarizes business concerns:
“Universal Cart doesn’t just colonize the bottom of the funnel. It colonizes everything from the first search query to the final checkout, without your customer ever landing on your site. This time, the adaptation window can be much shorter than a decade.”
The user these demos were designed for
After watching the I/O demos, it became clear that these features are aimed at a specific type of user. This user does not open ten tabs and compare options manually. They describe what they want and let AI do the rest.
When they ask information officers to monitor apartment listings or track sneaker drops, they’re not searching in the traditional sense. They delegate a search task and wait for a notification.
This means that companies are competing for something different. Haroon Qureshi, Global Head of Retail Experience and Partnerships at WPP Media, describes how goals have changedstating:
“In the future, will brands compete for clicks? Or to be recommended?”
At I/O, Google said AI Mode has surpassed 1 billion monthly users, with queries more than doubling each quarter since its launch. This gives this new way of searching a reach that few interfaces can match.
Why it matters for research professionals
E-commerce
Google guarantees that with UCP, your brand remains the merchant of record. Shoppers can pay with Google Pay or transfer items directly to your business website.
However, marketers are starting to distinguish between owning the purchase and owning the data that led to it.
Armando Roggio, main collaborator of Convenient e-commerceput it directly:
“In Google’s model, merchants still own the transaction, but not the purchase intent or product discovery.”
This makes the optimization problem harder to solve without Google’s data on how different signals are weighted in agent-mediated feeds.
Aleyda Solís, SEO consultant and founder of Orainti, noted on LinkedIn that “ecommerce SEO and AI search optimization cannot be reduced to “product content.”
His article describes important signals, such as accurate feeds, consistent attributes, clear pricing, and detailed content that gives agents something to reason about.
Local and service businesses
For local businesses, Search aggregates pricing and availability with direct links to complete the booking through the provider of your choice. In certain categories such as home repair, beauty and pet care, users can ask Google to call businesses on their behalf.
If the call goes to voicemail or if staff can’t provide clear answers, the company risks losing the moment before the user visits a site.
In a way, the booking agent transforms preparation into a visibility factor. Karim Al Chamaa, founder of Implent, described the dynamic on his company blogstating:
“When it’s Google’s agent calling, disorganization automatically becomes disqualification.”
Measures
If an information agent monitors apartment listings for a week and returns a recommendation, the value has been extracted without a conventional click path.
Jake Ward, co-founder of Mentions, posted on that “we are moving further into a world of visibility > clicks”. You can track organic sessions and referral clicks, but you can’t track how often your business’s products were considered and rejected by an agent, or how often your business was recommended in an agent booking flow.
Metrics that have explained search performance for years may not explain these agent-mediated pathways as clearly.
What we don’t know yet
Google has not shared the selection criteria for Universal Cart recommendations or agent booking results. Marketers currently develop strategies based on inferences rather than official guidelines. Until Google clarifies the signals its agents rely on for comparison and selection, the optimization process remains a matter of considered guesswork.
Currently, there are no third-party measurement tools that track agent-initiated transactions or the frequency with which recommendations are made as separate metrics from organic traffic.
While Merchant Center now provides AI-powered insights that compare share of voice against similar brands, businesses can’t tell if “the agent never considered us” or “the agent considered us and rejected us.”
The connection between paid ads and organic viewability in AI-driven commerce is also not fully explained. Google mentions that it is “neither a retailer” nor “a marketplace,” but Universal Cart aggregates products from various merchants and offers AI feedback suggesting alternatives. How advertising fits into that experience is a question Google hasn’t answered.
Looking to the future
Google allows consumers to move from research to action more quickly, but at the same time, it makes it harder for businesses to see and measure their visibility. The shared shopping experience at I/O was presented from the consumer side, with few details provided that could help businesses appear there.
The feedback loop becomes increasingly difficult to follow. When a consumer leaves the purchasing decision to an agent, businesses that weren’t chosen may never know they were part of the process.
More resources:
Featured Image: Romain Samborskyi/Shutterstock
