Amazon's Rufus AI assistant gains memory, price tracking and auto-buying

Amazon announced on November 18, 2025, that its AI shopping assistant Rufus has received more than 50 technical upgrades since its initial launch, transforming the tool into what the company describes as a personalized shopping companion capable of autonomous purchasing and comprehensive price tracking.

More than 250 million customers used Rufus this year, with monthly average users increasing 149% and interactions climbing 210% over the past year, according to the announcement. Customers who engage with the assistant while shopping demonstrate 60% higher purchase likelihood during those sessions.

The enhanced system now incorporates account memory that understands individual shopping patterns. A customer who previously mentioned having two sports-enthusiast children and a golden retriever will receive recommendations that prioritize age-appropriate athletic gear and vacuum cleaners optimized for pet hair removal. The memory extends beyond product preferences to encompass browsing history, past purchases, and stated household characteristics.

"We've built and keep improving our next-gen in-store AI assistant to be a knowledgeable shopping expert right at your fingertips—one that knows you and your preferences, can help you discover products, compare options and prices, make informed buying decisions, and even take action to get things done faster for you," said Rajiv Mehta, vice president of Search and Conversational Shopping at Amazon.

The system operates through Amazon Bedrock, utilizing multiple large language models including Anthropic's Claude Sonnet, Amazon Nova, and a custom model trained on Amazon's product catalog. A real-time router directs queries to appropriate models based on capability requirements, latency constraints, and answer quality optimization. The architecture employs Retrieval-Augmented Generation to incorporate insights from publications including The New York Times, USA Today, Good Housekeeping, and Vogue when addressing product questions and trend analysis.

Autonomous purchasing capabilities

Rufus now executes purchases without direct customer intervention through several mechanisms. Customers can request reorders using natural language commands such as "Reorder everything we used to make pumpkin pie last week" or "Order the hiking boots and poles I browsed yesterday." The system connects past activity to current needs, suggests alternatives when items become unavailable, and automatically adds selections to shopping carts for review before checkout completion.

Price tracking represents another autonomous feature. The assistant displays 30-day and 90-day historical pricing data, enabling customers to evaluate deal quality immediately. Prime members can establish price alerts for specific products, with the system monitoring prices continuously on their behalf. When target price points arrive, Rufus completes purchases agentically using default payment methods and shipping addresses, sending notifications with 24-hour free cancellation windows. Auto-buy requests remain active for six months or until cancellation, with customers averaging 20% savings per purchase through this feature.

The autonomous purchasing functionality positions Amazon directly against third-party AI shopping agents that the company has blocked through legal action. Amazon filed a federal lawsuit against Perplexity AI on November 4, 2025, alleging unauthorized deployment of AI agents into the e-commerce platform. The company maintains that controlling the AI shopping experience protects its $56 billion advertising business built around customers browsing its marketplace.

Visual and text search expansion

The updated system introduces activity-based product discovery. Customers can search for items needed for specific occasions—hosting a Frozen-themed birthday party, spending a day golfing, or preparing third-grade school supplies. Depending on query specificity, Rufus either suggests shoppable product categories for browsing or recommends specific items it can add directly to carts.

Visual search capabilities expanded for iOS users through handwritten list recognition. Customers can photograph grocery or shopping lists, upload images, and Rufus adds items directly into Amazon carts, with Android functionality scheduled for future release. Upcoming features will enable users to upload photos of problems requiring solutions—a stained rug, for example—and receive tailored cleaning supply recommendations based on fabric analysis.

The assistant can locate influencer storefronts through simple requests like "Find Paige DeSorbo's storefront." It also discovers products not currently available in Amazon's inventory. When customers search for branded items unavailable on Amazon, Rufus displays either a "Buy for Me" button enabling Amazon to purchase from third-party websites on their behalf, or a "Shop Direct" button redirecting to merchant websites.

This "Buy for Me" feature reflects Amazon's broader strategy of maintaining control over AI-mediated shopping experiences. The company blocked AI bots from OpenAI, Anthropic, Meta, Google, and Huawei on August 21, 2025, while simultaneously developing internal tools that compete with external AI agents.

Budget-conscious shopping features

Rufus functions as what Amazon characterizes as a "smart deal finder" capable of curating personalized promotions throughout the year, including Prime Day, Black Friday, and Cyber Monday. Customers can search by price constraints using queries like "Men's jeans under $40" or "Best toys under $15" to filter selections matching budget parameters.

The personalized deal curation draws from browsing patterns, shopping cart contents, and wish list activity. This approach contrasts with traditional promotional displays that present identical offers to all customers regardless of individual preferences or purchase history.

For marketing professionals tracking AI's impact on advertising, Amazon's third quarter results demonstrated Rufus's commercial effectiveness. The company reported that customers engaging with the AI assistant complete 60% more purchases during shopping sessions. Amazon's advertising revenue reached $17.7 billion in Q3 2025, growing 22% year-over-year, significantly outpacing overall sales growth of 12%.

Customer service integration

The assistant operates as a 24/7 customer service system handling package tracking, delivery status inquiries, return and refund process explanations, replacement guidance for damaged items, order modifications, cancellation assistance, account settings navigation, payment and billing questions, login and security support, and connections to live customer service representatives when necessary.

This customer service functionality extends Amazon's agentic AI deployment beyond consumer-facing applications. The company introduced autonomous AI capabilities for marketplace sellers on September 17, 2025, enabling automated inventory management, compliance monitoring, and advertising campaign development through systems that operate without constant human oversight.

Technical infrastructure

The system's intelligence upgrades encompass general shopping knowledge, category and product research methodologies, evaluation frameworks, and search and recommendation algorithms. Amazon has expanded the underlying neural architecture while building a knowledge graph of product attributes and customer preferences.

Rufus appears throughout Amazon's ecosystem: the homepage, product detail pages, Amazon Shopping app, desktop interface, and Amazon Lens Live experience. Customers initiate conversations by typing natural language queries into the assistant's chat interface or speaking through microphone input.

The coming months will bring memory expansion to encompass customer activity across Amazon's digital services including Kindle, Prime Video, and Audible. This cross-service memory integration aims to provide more comprehensive personalization based on entertainment consumption patterns alongside shopping behavior.

Industry implications

The Rufus enhancements arrive as major retailers adopt divergent AI agent strategies. Research published in September 2025 found that most merchants welcome AI-mediated shopping experiences while maintaining control over customer interactions. Amazon represents the most significant exception to merchant openness, explicitly blocking external AI bots through robots.txt configurations while developing proprietary alternatives.

Industry analysts view Amazon's approach through competitive lenses. Eric Seufert from Mobile Dev Memo argued that blocking AI agents serves Amazon's business interests because agentic commerce "violates the motivations of retail outlets to control the customer relationship and monetize their first-party data with advertising."

Google launched competing autonomous shopping capabilities on November 13, 2025, including agentic checkout that automatically purchases tracked items when prices decline, Duplex-powered phone calls to verify local inventory, and conversational shopping through AI Mode and Gemini app. These features mark a substantial expansion of Google's Shopping Graph infrastructure containing more than 50 billion product listings.

The shift toward AI-mediated commerce creates measurement challenges for marketing professionals. When autonomous agents control shopping experiences, traditional advertising models relying on visual impressions and clicks face structural disruptions. Advertisers must adapt strategies for environments where AI assistants prioritize user intentions over platform monetization objectives.

Amazon's development of Rufus reflects broader patterns in advertising technology. The company unified its DSP and sponsored ads console on November 10, 2025, launching AI agents for campaign management while extending Creative Agent capabilities to television formats. These announcements arrived alongside closed beta access to a Model Context Protocol Server enabling natural language interactions with advertising APIs.

However, skepticism persists regarding AI shopping agent viability. Analysis published in October 2025 identified eight structural challenges facing autonomous shopping systems, including retailer incentives against AI intermediation, consumer preferences for seeing options before deciding, and AI agents lacking contextual purchase knowledge that influences product selection.

The competitive dynamics intensified when Amazon filed lawsuit against Perplexity over the startup's Comet browser on November 4, 2025. The complaint alleged that Perplexity deployed covert AI agents that disguised themselves as Google Chrome while enabling purchases through customer accounts. Amazon characterized this as degrading the shopping experience the e-commerce platform spent billions developing.

Timeline

• February 1, 2024: Amazon unveils Rufus conversational AI shopping assistant in beta testing
• July 12, 2024: Amazon expands Rufus to all US customers ahead of Prime Day 2024
• March 26, 2025: Amazon launches Interests feature for personalized shopping prompts
• August 21, 2025: Amazon blocks AI bots from major tech companies including OpenAI, Anthropic, Meta, and Google
• September 2, 2025: Amazon launches Lens Live with real-time scanning and Rufus integration
• September 17, 2025: Amazon introduces agentic AI across seller platform for autonomous business management
• October 27, 2025: Amazon launches Help Me Decide feature using AI to recommend products
• November 4, 2025: Amazon files lawsuit against Perplexity over unauthorized AI agent deployment
• November 18, 2025: Amazon announces 50+ technical upgrades to Rufus including memory, price tracking, and autonomous purchasing

Summary

Who: Amazon announced the enhancements, affecting more than 250 million customers who used Rufus this year. Rajiv Mehta, vice president of Search and Conversational Shopping at Amazon, provided statements about the system's capabilities.

What: Amazon deployed over 50 technical upgrades to Rufus including account memory that learns individual shopping patterns, autonomous purchasing through reorder commands and price-triggered auto-buying, 30-day and 90-day price tracking with alerts for Prime members, visual search capabilities including handwritten list recognition, activity-based product discovery, and "Buy for Me" features for third-party purchases. The system operates through Amazon Bedrock using multiple large language models including Anthropic's Claude Sonnet and Amazon Nova.

When: Amazon announced the enhancements on November 18, 2025. Monthly average users increased 149% and interactions climbed 210% over the past year. More than 250 million customers used Rufus this year.

Where: Rufus operates throughout Amazon's ecosystem including the homepage, product detail pages, Amazon Shopping app, desktop interface, and Amazon Lens Live experience. The assistant is available to customers in the United States, with iOS users receiving handwritten list recognition and Android functionality scheduled for future release.

Why: The enhancements aim to transform Rufus from a basic question-answering tool into a personalized shopping companion that saves customers time and money. Amazon positions the technology to compete against third-party AI shopping agents while protecting its $56 billion advertising business. Customers who use Rufus demonstrate 60% higher purchase likelihood during shopping sessions, with auto-buy features delivering average savings of 20% per purchase.