Field Data No. 01

The Rise of AI in Ecommerce Search

Google spent 2026 building a place to shop inside Google. Here is the sequence, with the dates, and what it means if you sell things online.

On January 11, 2026, at the National Retail Federation’s annual event, Google announced the launch of the Universal Commerce Protocol (UCP), a “new open standard for agentic commerce and AI tools to help retailers connect with high‑intent shoppers and drive sales.” A month prior to this, a Google core algorithm update began setting the stage for what is coming because of this launch.

Since 2024, we have seen AI Overviews grow to dominate the top of search results for informational queries. But prior to and shortly following the launch of UCP we have seen a dramatic increase in the number of AI Overviews appearing in commercial queries, which we evaluate by measuring AIO citations in which a product page is the cited page.

Fig. 1Keywords returning an AI Overview citation
1K 10K 100K 1M 10M JUL 25 NOV 25 MAR 26 JUL 26 11 DEC CORE UPDATE AMAZON WALMART HOME DEPOT TARGET BEST BUY CHEWY 364× BEST BUY 3.8× AMAZON 2.6× WALMART 2.2× HOME DEPOT 1.8× TARGET 1.7× CHEWY
Six retailers, July 2025 to July 2026, on a log scale — Amazon holds more than three times the other five combined, so a linear axis would flatten them against the floor. Four of the six peak in April 2026.
  • Amazon3.8× increase in citations
  • Best Buy364× increase in citations
  • Walmart2.6× increase in citations
  • Home Depot2.2× increase in citations
  • Target1.8× increase in citations
  • Chewy1.7× increase in citations

Along with testing the performance of agentic commerce in the wild, Google has been working to establish two new habits for searchers throughout 2026.

First, searchers are being trained to expect zero‑click product breakdowns powered by AI. Whereas traditionally one might visit a few retail sites and read reviews from independent publishers, now they get the pros, cons, and specifications in search.

Second, they are being trained to trust Google’s AI as the primary curator — an advisor that gives them the breakdowns and comparisons they need to make a choice.

Section OneGoogle makes itself the answer

Another launch came on May 19, 2026. This was Google’s Universal Cart, unveiled at Google I/O. With it, Google claimed to be “building the foundation for agentic commerce.”

Two days after this announcement, another core algorithm update was initiated, and with it another change to the ecommerce search landscape.

With the rollout of this core algorithm update, Google began taking up share of AIO citations itself.

Fig. 2Google’s own share of AI Overview citations
0% 3% 6% 9% 12% JAN FEB MAR APR MAY JUN JULY 21 MAY CORE UPDATE 10.1%0.8% GOOGLE.COM SHARE OF AI OVERVIEW CITATIONS, 2026
Flat between 0.24% and 0.78% for four months, then 1.26% in May, 7.66% in June, 10.09% in July. The rule marks the day the May core update began rolling out.

The gain is not quite what it first appears. YouTube gave up five points over the same window, and YouTube is Google. Combine the two and the picture is a straight exchange.

Fig. 3A one-for-one swap
0% 13% 26% 39% 52% 65% JAN FEB MAR APR MAY JUN JULY 59.9%48.8% INDEPENDENT SITES8.3%19.0% GOOGLE-OWNED GOOGLE.COM + YOUTUBE vs. ALL OTHER DOMAINS · −11.1 OUT, +10.7 IN
Google-owned properties gained 10.7 points; every other domain on the web lost 11.1. Nobody had to be pushed out for independent publishers to lose share — the pie simply stopped growing first.

When the Google citation is clicked, the shopping panel for that product opens. Notably, if the product is sold by one of Google’s UCP partners, a “Buy” button appears that enables you to buy the product from that retailer without leaving Google search.

As of this writing, I have witnessed several commercial AIOs that give Google platforms near exclusivity in the results between its shopping experience and YouTube, with perhaps one other citation.

We have seen Google in recent months both serving as the marketplace and competing for share of search with the very retailers selling on its marketplace. In my opinion, this is not the ideal state for Google, but another milestone in a larger effort.

To understand where we are headed with this, it is important to know what Google’s UCP is.

Section TwoWhat is Universal Commerce Protocol?

At the strategic level, UCP enables Google to manage every aspect of an online shopping journey on its own platform. A shopper still buys from Best Buy, but they do not need to see Best Buy’s website. This is not the first time Google has launched a project like this.

In 2018, Google launched “Shopping Actions,” often called Buy on Google, with the same objective. Retailers like Target and Walmart initially supported the effort, seeking ways to counter the growing dominance of Amazon. But this support waned because of who held the power, and how the technology was built. Buy on Google required retailers to adapt to Google’s platform rules, and Google shielded email addresses and loyalty details.

Buy on Google was also widely rejected by consumers. At this point, Amazon was the place to shop and Google was the place to search. Google trying to get into the online store business was a hard sell, the marketing was poor, and Google ultimately shut it down in 2023.

With UCP, Google makes buying on Google technologically easy. It is not a pitch to consumers to try a new Google platform, it is a functionality within Google search. But aside from making it seamless, Google will still face the same hurdle of adoption. Will making it easier to purchase on Google lead to people changing habits? Time will tell.

As for the retailers, with UCP they are still the merchant of record, and their brand gets full visibility. There is an API connection that allows cart logic, loyalty programs and user profiles to sync, so customer data is passed to the retailer. While retailers still provide a base product feed to Google Merchant Center for discovery, UCP replaces static website scraping by allowing Google’s AI to communicate directly with the retailer’s backend to pull live inventory, local stock, and final pricing in real time.

Section ThreeWhy is Google doing this?

As is often the case with Google, I think there is an ostensibly user‑friendly reason for this, along with overriding commercial reasons. I will start with the former.

Reducing shoppers’ friction caused by retailer websites

How long has Google been asking websites to improve their user experience? Core Web Vitals were established in 2020 to show companies just how badly their websites performed, but for anyone who has ever worked inside an enterprise ecommerce business, most development teams view those goals as nearly impossible to meet given the complexity and third‑party bloat of the sites they are working on. Furthermore, product teams are consistently making updates without considering the impact on load times.

Now, Google does not have to worry about your web performance leading to inefficient ad spending or unhappy searchers. They get control over the experience themselves.

Maximizing ad impression opportunities

Every time a searcher clicks to go to the retailer’s site, Google loses control of that interaction. Once the customer lands on the retailer’s domain, the retailer has control of advertising, cross‑selling, and purchase behavior. By keeping searchers within Search, AI Mode, or Gemini, Google ensures that the searcher stays inside its own ecosystem for the entire journey, with more opportunities to serve ads.

Getting more data on behavioral intent

With the shopping journey and checkout contained within Google’s ecosystem, they can learn more about a searcher’s intent. They see what was searched for, how options were evaluated, and what was ultimately purchased. Before the dawn of UCP, this data would have to be largely inferred, and some of it was lost once the customer was on the retailer site.

This data is the driver of Google’s largest revenue engine, search advertising. Complete intent data coupled with deterministic, closed‑loop conversion tracking allows Google to maximize ad targeting efficiency and definitively prove attribution. By proving higher ad effectiveness, Google builds a compelling case for retailers to increase their Google ad budgets.

Section FourWhat comes next?

Let us first review where we are now:

  • In January of 2026, Google began heavily introducing AI Overviews into search results with shopping intent.
  • In May of 2026, Google began pushing its own properties — Google and YouTube — into the citations of these AI Overviews, with the ability to buy inside search.

When I look at Google’s AIO citations, I do not see a finished product. I cannot think of a UX designer being satisfied with a small link from Google to Google that opens an already established Google result type.

What I do see is the plumbing for a major shift in Google’s shopping‑intent results in the future. Google already made the move in 2024 to make itself the product listing page by inserting Popular Products grids. A natural next step would be to make the product details page itself as well.

Look at the Product Knowledge Panel. All of the fundamental pieces of a PDP are already there: price, specifications, comparisons, images, videos, reviews. The only thing missing is better design.

With that shift, Google would become as close as one can get to Amazon without being Amazon.

Section FiveWhat this means for ecommerce marketers

Complete and accurate product data is critically important

I still maintain that most of what people in this industry call AEO or GEO requires many of the same actions as old‑fashioned SEO. For professionals in category management, merchandising, and SEO, the time for change is immediate. Websites that do not have a Merchant Center data feed containing full, rich specifications will not fare well.

Tighter competition for organic results

The middle class of organic search results is being squeezed. Regardless of how far down the search results people scrolled, blue links gave small and mid‑sized retailers the opportunity to earn clicks. I think it is unclear what will happen to the blue links, but Google’s AI recommendations already appear to favor the big box and major ecommerce players, because those giants have massive brand entity authority and perfectly optimized API and feed infrastructure.

More advertising spend

Shrinking real estate and improving Google’s own ad data both lean towards more advertising spend. Again, this has always been the case — if you want to be at the top of search, you must pay. But a more competitive layout and better understanding of attribution mean you are going to have to pay more.

Brand is more important than ever

At the end of the day, before customers make a decision on Google, they make a decision to use Google. This is the point Amazon attacked in 2025 when they pulled their product listing ad spend from Google. They are betting that they do not need Google because people will come to them first anyway.

Retailers need to give customers reasons to shop with them without having to use Google Search.

About this series

Field Data is one chart from real accounts, and the argument it supports. Sources and method are stated on every figure, including the parts that undercut the point.