Google is extending generative AI beyond general search questions and into travel, shopping and clothing discovery. The new features combine information from across the web and give users more ways to narrow their choices inside Google’s search experience.
Travel searches draw on multiple sources
Google’s Search Generative Experience, or SGE, had recently begun rolling out to select testers in the U.S. Rather than simply returning relevant websites, it generates an AI answer, follow-up questions and selected human-written sources.
Now the experience is being extended to travel searches. A search for a location can produce a response that draws on photos, user reviews and websites, bringing several kinds of information together in one place.
That approach builds on Google’s accumulated location and user data. The search result can offer a broad picture of a destination while still connecting people to the sources behind the answer.
Shopping results add summaries and follow-up questions
Product searches bring together images, prices and links to stores, with AI-generated reviews composed from reviews found around the web. The page can also offer a generated list of factors to consider when buying a product. For Bluetooth speakers, for instance, battery life is one consideration.
Users can continue the search through questions linked to Google’s chatbot. Someone looking for a speaker could refine the initial selection by asking for the best-rated waterproof option. This makes the search more conversational: the first result can be a starting point, followed by narrower requests.
For clothing, AI also supports searches that combine criteria such as price, color and pattern. Instead of relying on a single broad query, shoppers can describe several attributes together to find apparel that fits their preferences.
Virtual models show clothing across sizes
A separate feature lets people view clothing on AI-generated models chosen to better reflect their own body shape and size. The aim is to help shoppers anticipate how an item might look on them, rather than judging it only from typical catalog models.
The virtual models cover sizes XXS to 4XL. The feature is initially available for women’s tops from select brands, while a version for men’s tops is expected later this year.
Google says the models are created with diffusion technology, the approach also used in image systems such as Midjourney and Stable Diffusion. The image model was trained with data from Google’s Shopping Graph, which brings together information from websites, prices, reviews, videos and product data provided directly by brands and retailers.
The model is called “TryonDiffusion.” Its science and an interactive demonstration are available on a project page on Github. In tests with human evaluators, they preferred TryOnDiffusion’s results by about 90 percent over existing methods.
More useful answers could keep shoppers in Google
These features put more of the search journey inside Google. People can explore destinations, compare product information, get purchase considerations and refine what they want through follow-up questions. For clothing, generated models and attribute-based searches add another layer to product discovery.
The broader effect may be to strengthen Google’s position in search. If users can do more of their research within its AI services, they may spend more time in Google’s ecosystem. That in turn creates more room for advertising, as the source article observes.
The expansion therefore combines practical search tools with a larger shift in how results are presented. Google is using generative AI to organize existing information, summarize opinions and support visual product exploration, while keeping the next question and the next search close at hand.