Amazon weighs hybrid AI search on main retail site
Updated
Updated · The Information · May 5
Amazon weighs hybrid AI search on main retail site
13 articles · Updated · The Information · May 5
Vice-president Amanda Doerr said Amazon may add conversational blurbs above some search results, after tests spotted on Monday, while keeping Rufus separate for many queries.
Amazon says standard search suits direct purchases, while Rufus helps exploratory shopping, so any integration would be gradual and vary by task rather than replace the main search bar.
The move could affect Amazon's $68.6 billion advertising business; the company says Rufus drove nearly $12 billion in annualised incremental sales in 2025.
With AI like Rufus making shopping decisions for us, are we trading personal choice for ultimate convenience?
As Amazon's AI shops across the web for you, can other retailers survive being just fulfillment centers in its ecosystem?
When an AI assistant becomes the primary shopper, how can brands ensure their products are the ones being recommended?
Scaling Hybrid AI Search at Amazon: 250 Million Users, 60% Higher Purchase Rates, and Multi-Domain AI Integration
Overview
Between 2025 and 2026, Amazon transformed its search capabilities by upgrading OpenSearch with GPU acceleration and batch inference, significantly improving speed and accuracy. This infrastructure supported the launch of the Nova AI model, which combines semantic understanding with precise filtering for faster, deeper search results. Enhancements to the Rufus shopping assistant, including memory and auto-buying features, drove massive user growth and increased purchase rates by 60%, creating strong demand for Amazon's hybrid AI system. This integrated approach, combining advanced AI models with personalized agents, enabled Amazon to build a seamless, data-rich shopping experience and establish multiple growth loops across AWS, advertising, and commerce.