Here's what Google's AI-powered search looks like

Google's announcement of new features for its generative AI search experiment reveals telling details about the company's vision for the future of its core product.

On Wednesday, Google introduced new capabilities for what it's calling Search Generative Experience (SGE), highlighting travel and shopping as use cases. Compared to the product that was demo-ed at Google I/O in May, results now show information like user reviews, images, and product information in the snapshot at the top of the page. Rather than scrolling down the page and clicking on external links, users get all of the information they need in one condensed format. This is helpful for users, but potentially disastrous for publishers that rely on search engine traffic.

The vast majority of Google users haven't tried SGE yet, because it's currently limited to users with access to Labs, a testing ground for early-stage experiments. But today's announcement presents one of the first examples of SGE in action, highlighting some of the ways traditional search could be upended.

In one demo, results for "what do people say about the bean in Chicago? Is it worth visiting?" show a description of the famous Cloud Gate sculpture, aka "the Bean" accompanied by articles referenced by the response and user reviews relevant to whether the landmark is worth the hype. The snapshot takes up most of the page, which is normally occupied by search results.

In a demo of searching for the best bluetooth speaker for a pool party, Google showed results that pulled recommendations from articles found on the internet with information about where to buy it.

While this isn't shown in the demo, but a Google spokesperson said when you click on one of the speakers, it opens up a panel that shows you different prices from retailers and top reviews and considerations about the products. You can then click through to the retail site to buy it.

The spokesperson said Google is prioritizing approaches that send traffic to publishers, by including relevant content within the snapshot. But that also might make users less inclined to click into the article, thus losing out on potential affiliate revenue.

When asked how Google is taking the product's impact on publishers into consideration, Rany Ng, VP of Search said SGE leverages Google ranking systems to recommend publisher content normally found in search results so users can dive deeper into a query. "One of the frontal fundamental focus areas of designing SGE was to think about how we can make publisher content more more front and center."

Ng also emphasized that SGE is still an experiment and Google wants to hear feedback from both users and stakeholders, like advertisers and publishers.

Google has not revealed when these new features will be widely released. If you want to try the new features sooner than that, you can check out (and likely join the waitlist for) Google Labs.