Google tweaks its AI shopping experience ahead of the holidays

Google has made it easier for retailers to get discovered across its AI surfaces and added shopping features to YouTube to make it easier for people to buy directly while watching videos.

At its annual Rethink event in September, Google introduced a new AI performance insights tab within Merchant Center that will give retailers a look at what users are typing when shopping for products. This feature is unique because retailers will be able to see the language used by shoppers on Google’s AI surfaces like AI Mode, as opposed to synthetic prompts or guesswork offered by other tools on the market, Jason Goldberg, chief commerce strategy officer and Publicis Groupe, told Retail Brew.

“This AI performance insight is the first time we have real user data on how people are actually using the surface with AI to shop, and so that’s a very important new data signal,” Goldberg said.

The tech giant also started beta testing its business agents (or AI chatbots that customers can use to chat with brands), first announced in January for YouTube ads and asked retailers in the US to try it out. “This is what’s really emerging as a big trend at the moment is these new AI ad formats that are essentially conversations with an ad [or] chat ads,” Sky Canaves, eMarketer analyst for retail and e-commerce, told Retail Brew.

Overall, Google’s Rethink event was more of a progress update on the adoption of Google’s existing agentic shopping features, many of which were first announced at NRF in January, rather than a flood of new announcements. Google also shared the progress of pilots with different brands.

Search for sales: Tapestry, Google said, has recently started using UCP-powered checkout so shoppers can buy Coach and Kate Spade products directly from Google’s AI search results, including AI Mode and on Gemini, for the upcoming holiday season.

Google also said merchants using UCP-powered checkout can now transfer carts to their own sites, test different checkout flows, view analytics on performance, and access new features rolling out soon.

While Google continues to invest in AI shopping, Canaves said, adoption of Google’s tools remains “fairly limited.”

“The question is, what will get consumers to adopt it now when they didn’t really adopt it earlier?” Canaves said. “Is Google going to push more of that on them? Are there going to be significant incentives? Google still identifies itself mainly as a matchmaker [between buyers and sellers], and not as a marketplace or as a shopping destination.”

However, Goldberg suggested that the large base of Google search users could potentially convert to shopping:

“They have an unmatched ability to put these AI experiences in front of a huge amount of consumers that already have some level of purchase intent,” he said. “Google has a chance to win, and I think Google has been very thoughtful about not just rushing features out, but understanding the technical infrastructure required to make them successful in the long run.”