๐ฒDigital TrendsโขStalecollected in 37m
Amazon rolls out eight new AI-powered visual search features
๐กSee how Amazon is integrating real-time generative AI into visual search to redefine the e-commerce shopping experience.
โก 30-Second TL;DR
What Changed
Introduced eight new visual search features including Lens Live and Circle to Search.
Why It Matters
This update signals Amazon's aggressive integration of multimodal AI into the e-commerce funnel, potentially increasing conversion rates by reducing friction in product searching. It sets a new standard for retail apps competing with Google's visual search dominance.
What To Do Next
Analyze how Amazon's visual search UX patterns can be adapted for your own e-commerce or inventory management AI applications.
Who should care:Developers & AI Engineers
Key Points
- โขIntroduced eight new visual search features including Lens Live and Circle to Search.
- โขFeatures real-time AI-generated images directly within the search bar.
- โขFull availability across both iOS and Android mobile applications.
- โขFocuses on improving user intent recognition through visual inputs.
๐ง Deep Insight
Web-grounded analysis with 17 cited sources.
๐ Enhanced Key Takeaways
- โขLens Live integrates Amazon's AI shopping assistant, Rufus, to offer product insights, summaries, and answer questions directly within the live camera view, enhancing product research.
- โขThe 'Circle to Search' feature specifically allows users to isolate and search for particular items within an image that contains multiple products, providing more precise visual search capabilities.
- โขOne of the new visual search features generates AI images in real-time within the search bar as customers type descriptive language, helping them visualize and discover products that match their textual descriptions.
- โขLens Live employs an accurate, lightweight computer vision object detection model running on-device, coupled with a deep learning visual embedding model, to identify and match products against billions of Amazon items in real time.
- โขUsers can now instantly launch the Amazon Lens Camera Search directly from their iOS lock screen using a dedicated widget, streamlining access to visual product discovery.
๐ Competitor Analysisโธ Show
Competitor Analysis: Visual Search in E-commerce
| Feature / Platform | Amazon (Lens Live, Circle to Search) | Google (Lens, Circle to Search) | Syte (E-commerce Visual AI) |
|---|---|---|---|
| Core Functionality | Real-time visual product search, AI-generated image search from text, multi-object selection. | Universal visual search across apps, multi-object selection, text + image search. | Visual AI-powered search for fashion/home goods, image-based product discovery. |
| AI Assistant Integration | Integrated with Amazon's Rufus AI shopping assistant for product insights and Q&A. | Integrated with Google's AI for enhanced search results. | Focus on visual matching, less emphasis on conversational AI assistant. |
| Real-time Processing | Lens Live offers continuous real-time scanning and instant matches. | Google Lens offers live camera feed search; Circle to Search is real-time on-screen. | Primarily image-based product discovery, real-time aspects depend on implementation. |
| Multi-Object Selection | 'Circle to Search' allows users to specify items within an image. | 'Circle to Search' allows circling, highlighting, scribbling, or tapping on-screen elements. | Focuses on highly accurate visual matching for specific items. |
| Text + Image Search | Allows refining visual searches with text qualifiers. | Supports multi-search with text and pictures simultaneously. | Can combine visual input with textual attributes for refinement. |
| AI-Generated Images (from text) | Generates AI images in real-time as users type descriptive search terms. | Not a primary advertised feature for direct product generation from text in search. | |
| Platform Availability | Amazon Shopping app on iOS and Android. | Select premium Android smartphones (Pixel 8, Galaxy S24 series) for Circle to Search; Google Lens widely available. | E-commerce platforms, often integrated as a backend solution. |
| Pricing / Benchmarks | N/A (Integrated into Amazon shopping experience) | N/A (Integrated into Google's search ecosystem) | Enterprise-level, custom pricing based on business size and catalog complexity. |
๐ ๏ธ Technical Deep Dive
- Lens Live leverages AWS-managed Amazon OpenSearch and Amazon SageMaker services for deploying machine learning models at scale.
- It utilizes an accurate, lightweight computer vision object detection model that runs directly on the user's device for real-time product identification.
- A deep learning visual embedding model is employed to match the customer's camera view against billions of products in Amazon's catalog, retrieving exact or highly similar items.
- For general visual search capabilities, Amazon's architecture can involve using the Amazon Titan Multimodal Embedding G1 model hosted on Amazon Bedrock to create vector embeddings for products.
- These vector embeddings are then used with k-Nearest Neighbor (KNN) search in Amazon OpenSearch Service to find semantically similar results, using Euclidean distance or cosine similarity.
๐ฎ Future ImplicationsAI analysis grounded in cited sources
Amazon will see increased conversion rates and customer engagement through enhanced product discovery.
By making product discovery more intuitive and immediate through visual and AI-generated search, customers are more likely to find desired items and complete purchases, reducing friction in the shopping journey.
There will be further integration of AI assistants into various stages of the shopping workflow.
The integration of Rufus into Lens Live suggests a strategic move towards conversational AI guiding visual search, providing product information, and assisting with purchasing decisions, extending beyond simple search results.
Amazon will gain richer data for personalization and recommendation engines through multimodal search inputs.
Combining diverse inputs like real-time visual scans, user-drawn selections, text descriptions, and interactions with AI-generated images provides a more comprehensive understanding of user intent, enabling more precise and personalized product recommendations.
โณ Timeline
1994
Amazon founded as an online bookstore.
2003
A9.com, Amazon's search engine subsidiary, launched.
2017
Amazon SageMaker, a fully managed service for building, training, and deploying machine learning models, launched.
2023-09
Amazon enhanced its visual search capabilities, allowing users to search for visually similar products using images and refine results with text qualifiers.
2025-09
Amazon introduced Lens Live, an AI-powered shopping tool that scans products in real-time and displays matching items in a swipeable carousel.
2026-06-03
Amazon rolls out eight new AI-powered visual search features, including Lens Live enhancements, Circle to Search, and real-time AI-generated images from text descriptions.
๐ Sources (17)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
๐ฐ
Weekly AI Recap
Read this week's curated digest of top AI events โ
๐Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: Digital Trends โ