Amazon Shopping App Launches AI-Generated Real-Time Product Images for Dynamic Virtual Previews and Search Assistance

On June 3, Amazon announced the addition of an AI real-time image generation feature to its shopping app’s search bar, now available to US users on iOS and Android, initially covering apparel and home categories. When users enter a descriptive search term, multiple sets of AI-generated reference product images will appear below the autocomplete suggestions, updating in real-time as each new word is added. Clicking the corresponding image will take users to the actual product results page that matches the visual style. The feature is designed to solve the pain point of users not knowing the exact terminology — for example, not knowing that a gathered neckline is called a “cowl neck” or a rattan side sofa is called “rattan” — helping users narrow down their search through visual guidance. The underlying technology is powered by Amazon Bedrock, including Claude Sonnet, Amazon Nova, and a custom model fine-tuned on Amazon product data.

It should be noted that these images display AI-generated imaginary products that do not actually exist on the platform, drawing skepticism from media outlets like TechCrunch. On May 13, Amazon renamed its shopping AI assistant Rufus to Alexa for Shopping, extending it to Amazon.com web and Echo devices, along with more than 50 technical optimizations. The timing of this launch is telling: Prime Day 2026 will be held from June 23 to 26, earlier than usual, while OpenAI, Google, Perplexity, and others are accelerating their AI shopping agent efforts. Analysts see Amazon’s move as a defensive play to protect its search entry point.

亿邦动力 | TechCrunch