Google Shopping added generative artificial intelligence to fashion listings this month, changing how some shoppers discover apparel items and reinforcing ecommerce fundamentals.
Shoppers are often of two minds, according to Google. Some have only a vague idea of what they want. Others have a clear vision.
“It can be hard to translate a vision for an item that fits your personal style (say, ‘colorful midi dress with big daisies’) into something you can buy and have in your closet by Friday,” wrote Lilian Rincon, a Google vice president, in a March 5, 2025 blog post.
Vision Match
Google Shopping’s new AI image generator aims to help shoppers find what they want. Called “Vision Match” in Google’s documentation, the feature is labeled “Create & Shop” on the customer-facing front end.
A shopper can type or speak a description, such as Rincon’s “colorful midi dress with big daisies.” The Vision Match AI generates images based on that description — flowered dresses in this case — and shares shoppable product listings similar to the generated images.
Vision Match ingests a text description and generates images such as the green dress with daisies shown here. Click image to enlarge.
Vision Match may function as a bridge spanning a shopper’s abstract idea and an actual product for sale.
Moreover, Vision Match pairs well with other Google features that deploy shopping data to improve ad performance and product discovery, including:
- Google Lens, which allows users to search for products by uploading images or taking photos.
- GenAI search in Google Shopping, such as tools that help shoppers find products.
- Google Shopping image search and style matching for fashion and home décor.
- Virtual try-on for beauty and apparel, allowing users to see how products look on models.
Improved Shopping
Google Shopping’s various AI tools will almost certainly improve consumers’ experiences. Folks use Google to shop more than a billion times a day, and the company has an excellent store of data.
Google knows what products are available via its Shopping Graph, which had 45 billion listings as of October 2024, as well as what shoppers want, e.g., a “colorful midi dress with big daisies.”
For example, the press kit Google’s media relations team shared with journalists ahead of the Vision Match announcement included a “trends” document that stated:
- “Cheetah print jeans” and “leopard jeans” are the top trending types of jeans.
- In April 2024, search interest in “baggy jeans” surpassed that of “skinny jeans” for the first time, and “baggy jeans” have remained on top ever since.
- “Shell skirt” is at an all-time high for the second consecutive month.
- Idaho is the only U.S. state where purple lipstick is the most popular.
For better or worse, Google knows much about shoppers (and advertisers). Google Shopping can find the needle in a haystack of 45 billion products.
Optimizing for AI
With Vision Match, Google is not reinventing ecommerce but becoming better at using the data.
Optimizing products for Google’s AI features typically includes:
- Aligning product listings for AI. Vision Match and other AI features use data from the Shopping Graph.
- Creating superior product descriptions. Describe the product’s physical specs and primary benefits.
- Using quality images. AI tools analyze product images for colors, features, and more.
- Advertising. Use Performance Max campaigns to ensure products appear across Google Shopping, Search, and YouTube.
None of these tactics, however, are novel. They are fundamental to selling products online. Since 1995 — the year Amazon and eBay launched — sellers have needed structured, descriptive, and visual product information promoted by advertising.
Thus Google Shopping’s AI initiatives are, in a sense, sensible ecommerce practices and an opportunity for merchants. What has worked well — an online seller’s existing tactics — is the path to success in an AI-driven future.