Amazon launches AI search that generates product images from text

Amazon has introduced an AI-powered search tool that generates product images on the fly as customers type descriptive phrases into its Shopping app. The feature, now rolling out to U.S. users on iOS and Android, aims to solve a common e-commerce friction point: finding items when you know what you want visually but lack the precise terminology.

The tool launches first in apparel and home categories, where visual details often determine purchase decisions. Instead of scrolling through generic results for "blue chair," shoppers can now type "blue velvet accent chair with gold legs and curved back" and watch AI-generated images update with each added detail.

How the technology works

The feature appears in the search suggestions area as users type. Descriptive queries trigger AI-generated images that evolve as the description becomes more specific. When a generated image matches the shopper's mental picture, tapping it surfaces visually similar products available for purchase.

Crucially, the AI-generated images themselves don't represent actual listings. They function as visual interpretation tools that help Amazon's algorithm understand style intent and match it to inventory. The system then displays real products with comparable characteristics.

Why it matters

This represents a fundamental shift in how search engines interpret shopping intent. Traditional keyword-based search requires users to know industry terminology—"tufted," "mid-century modern," "cap sleeve." Visual-first search lets customers describe what they see in their mind using everyday language. For furniture and fashion retailers, this could reduce the gap between browse and purchase, particularly for shoppers who think visually rather than categorically. The technology also signals how generative AI is moving from content creation into transaction facilitation.

The accuracy challenge

Amazon cautions that AI-generated images should be treated as sketches rather than product promises. The technology can create idealized visuals that may look better than available inventory, potentially creating a gap between expectation and reality. Shoppers should verify actual listing photos, dimensions, materials, and reviews before purchasing.

The feature fits into Amazon's broader visual search strategy. The company already offers Amazon Lens, which identifies products from camera photos, and Lens Live, which scans items in real time. Users can also combine image uploads with text modifiers—uploading a photo of a beige sofa and adding "in white" to refine results.

In apparel, Amazon has added "Shop by style" collages that use AI-generated imagery to group products around aesthetic themes like "urban luxe" or "soft elegance." These collections aim to provide outfit inspiration rather than single-item results.

The technology's effectiveness will depend on how well Amazon's visual matching algorithms perform after the initial AI generation. A polished generated image means little if the subsequent product recommendations miss key details that mattered to the shopper.

Details of the rollout were first reported by CyberGuy, which noted the feature is expanding gradually across product categories.