"To fashions" arrives in Italy: how the Google research integrated with artificial intelligence works

“To fashions” arrives in Italy: how the Google research integrated with artificial intelligence works

Google interface to fashions. Credit: Google

Googletoday 8 October 2025, has officially launched his new one in Italy To fashionsavailable directly in the Google search bar and within theGoogle app for Android and iOS. This novelty promises to transform the way we interact with the search engine, making it go from a simple list of links to a real smart assistantcapable of answering complex questions in a structured and conversational way. To fashions, present in almost 50 new countries and in 36 new languages, is able to provide organized summaries, comparisons and planning, integrating the generative artificial intelligence of Gemini directly on the research page. Giorgia Giulia Evangelista, Geopop’s Head of Content Development, had the honor of interviewing Nick FoxSr. Vice President, Google’s Knowledge & Information to deepen the ideas underlying this research model and its main characteristics such as accuracy, speed and reliability.

How it works to fashions: the technical engine

From a technical point of view, fashions are not limited to looking for keywords, but uses a personalized version of the models to the advanced google, Geminito understand the meaning of long and articulated questions. When a user asks a complex question, he uses it a technique called “Fan-out Query”: In practice, disappear the question in many sub-commodities, launches more research simultaneously and then assembles the results in a unique and coherent response. This allows you to explore the web deeper and to provide relevant answers that they would have requested numerous separate searches before.

The system is designed to be multimodalallowing you to ask questions not only through textbut also with the voice or loading Images directly in the search bar.

Search for Google Mode
How research with Google fashions appear. Credit: Google

Nick FoxSr. Vice President, Google’s Knowledge & Information, explained the idea behind Ai Fashus:

The main intuition, to create research based on AI AI most powerful we have ever built, was that users were looking for the power of research combined with the power of the AI. So we managed to combine the ability of thought, reasoning and the conversational nature of our Gemini models with the power of research, to provide great improvement.

What can be done with fashions: from travel itineraries to comparative tables

The goal of this new mode is to assist the user in homework that traditionally would take a long time. Here are some practical examples:

  • Plan complex activities: You can ask questions such as “organizes a three -day trip to Rome for a family with small children, including activities suitable for them and restaurants with seats”, and receive a structured itinerary.
  • Compare products: It is possible to ask for example, “I would like to understand the different methods of preparation of coffee. Create a table that compares the differences in taste, ease of use and the necessary equipment”.
  • Make insights: After receiving an answer, you can ask follow-up questions and the AI ​​will keep the context of the conversation.

Google’s model was optimized for theaccuracy el ‘reliability information and for the speed. Fox stressed that one of the things that leaves people openly is his rapiditythe answers arrive almost instantly.

Google stresses that the goal is to help people discover the contents of the web with a “superior quality

We also realized that the quality of our research information is really high. So the model was deeply optimized to do this.

How to use Google fashions

To fashions is available today as dedicated card which appears on the search results page and within theGoogle app for Android and iOS.

Google itself admits that, being a product in its initial stages, the results “may not always be perfect”. For research in which reliability is not considered high, the system will show the classic web results. The commitment, however, is to continuously improve the service on the basis of user feedback.