At the recent Google I/O 2025 conference, Google announced its new AI-driven project called the "Try On" experiment and has recently made it available to more users in the United States. This feature uses generative AI technology, allowing users to virtually try on clothing by uploading a photo, significantly enhancing the convenience and realism of online shopping. AIbase provides an in-depth analysis of the highlights of this innovative experiment and its potential impact on the e-commerce industry.
A New Virtual Try-On Experience: AI Makes Shopping More Intuitive
The Google "Try On" experiment is based on its latest image generation model, specifically designed for fashion scenarios, capable of accurately simulating the real effect of clothing on a user's body. Users simply need to select their favorite shirts, skirts, or pants on the Google Shopping platform, click the "Try On" button, and upload a full-body photo. The AI can generate realistic try-on results within seconds, showcasing the material, pleats, and fit of the clothing. This function relies on Google’s deep integration of data from 50 billion global items, ensuring that users receive real-time and reliable product information.
AIbase learned that the experiment currently supports virtual try-ons for skirts and pants, with plans to expand to more types of clothing, such as tops and outerwear, in the future. Compared to static images or model displays on traditional e-commerce platforms, the "Try On" experiment allows users to more intuitively determine whether clothing fits their body shape and style, greatly reducing hesitation costs in purchasing decisions.
Multi-Condition Search and Automated Shopping: Efficiency Upgraded
In addition to virtual try-ons, the Google "Try On" experiment also integrates multi-condition search and price comparison functions. Users can input complex requirements via natural language, such as “Find a blue dress suitable for formal occasions at a price below $100.” The AI will automatically filter matching options from the product database and provide price comparisons and real-time inventory information. More impressively, this feature supports an automated shopping process, allowing users to complete orders and payments directly through AI, eliminating cumbersome steps.
AIbase believes that this highly automated shopping experience not only improves user efficiency but also provides retailers with more precise consumer insights. For example, AI can recommend matching items based on user preferences or analyze popular clothing based on try-on data to support brands in optimizing inventory management.
The Secret Behind the Technology: Generative AI and Data Integration
The core of the Google "Try On" experiment lies in its customized image generation model, which has been optimized to precisely capture the texture, lighting, and fit of clothing. Combined with Google Search’s knowledge graph and real-time data processing capabilities, the AI can not only generate high-quality try-on images but also adjust effects based on users' body shapes and skin tones to ensure visual authenticity. Additionally, the experiment, supported by Gemini2.0, achieves multimodal query processing, allowing users to describe their needs through text, images, or voice descriptions to receive personalized shopping suggestions.
AIbase noticed that Google has put great effort into privacy protection. Personal photos uploaded are only used for generating try-on effects, and data processing strictly follows privacy compliance standards. Try-on images are neither stored nor used for other purposes, providing users with a secure environment.
Market Prospects: Reshaping Interaction Between E-Commerce and the Fashion Industry
Since its announcement at the Google I/O conference on May 20th, the "Try On" experiment has garnered significant attention in the U.S. market. AIbase learned that this feature is currently available for testing through Google Labs to U.S. users. Early adopters have highly praised its quick generation and realistic effects. Compared to Hume AI’s EVI3 and ElevenLabs’ Conversational AI2.0, Google’s "Try On" experiment focuses more deeply on optimizing the e-commerce scenario, demonstrating the unique value of AI in the consumer sector.
In the future, Google plans to expand the "Try On" feature to more countries and categories, and collaborate with retailers to introduce AR makeup try-on functions, such as virtual try-on and lipstick color matching. AIbase predicts that as the technology matures further, the "Try On" experiment could become a new standard in online shopping, driving the fashion e-commerce industry from "browsing" to "experiencing."
With its powerful AI image generation capabilities and seamless integrated shopping experience, the Google "Try On" experiment offers a win-win solution for both consumers and retailers. Its innovations in privacy protection, multi-condition search, and automated processes showcase Google’s ambitions in the AI-driven e-commerce field.