Amazon is turning Rufus into a persistent AI shopping companion that remembers your preferences, watches prices, and can even buy products on your behalf.
📌 Key Takeaways
- Rufus now runs across Amazon’s app and desktop with deeper personalization plus account memory.
- The assistant uses Amazon Bedrock, mixing Claude Sonnet, Amazon Nova, and a custom store model.
- It can auto fill carts from text, photos, and handwritten lists, then suggest alternatives if items are unavailable.
- New tools track 30 and 90 day price history, set alerts, and auto-buy at your target price.
- More than 250 million customers have used Rufus this year, with strong lifts in conversion and engagement.
Rufus Gets A Major Personalization Upgrade
Rufus now builds a profile around each shopper, using account memory to understand habits, household details, and product preferences over time. It learns from what you browse, buy, and explicitly tell it, then uses that context when it suggests products or answers questions.
Suppose you mention that you have two young kids, a shedding dog, or a preference for organic groceries.
In that case, Rufus will factor that into future searches and recommendations without you repeating it. You can also ask what it remembers, correct details, and add new information directly in the chat.
“We’ve built and keep improving our in-store AI assistant to be a knowledgeable shopping expert at your fingertips.” — Rajiv Mehta, Vice President, Search and Conversational Shopping, Amazon
Amazon plans to extend this memory beyond retail, so Rufus can eventually consider what you watch on Prime Video or read on Kindle when it suggests products, while still letting you manage what is stored.
The AI Stack Behind Amazon’s Shopping Assistant
Under the hood, Rufus routes each query across multiple large language models hosted on Amazon Bedrock.
It blends Anthropic’s Claude Sonnet, Amazon Nova, and a custom model trained on Amazon’s catalog, reviews, Q&A, and other store knowledge, choosing the right mix based on the task.
A real time router decides whether a question needs broad general knowledge, deep product detail, or fast search, balancing capability, latency, and answer quality for hundreds of millions of users.
Retrieval augmented generation then pulls in up-to-date insights from curated editorial and review sources to answer trend and product questions more reliably.
“Our goal is to save customers time and money by making online shopping even simpler.” — Rajiv Mehta, Vice President, Search and Conversational Shopping, Amazon
From Idea To Checkout In One Conversation
Rufus is designed to handle the full shopping journey, from vague ideas to final purchase, inside a single chat thread. You can ask broad questions like what you need for a themed birthday party or a weekend hiking trip, and the assistant will generate lists, suggest categories, and propose specific items.
The assistant now supports advanced visual search. Shoppers can upload handwritten grocery lists, which Rufus converts into carts, or share photos of problems, such as a stained rug, to get tailored product suggestions and step-by-step guidance.
It also connects to influencer storefronts and even products outside Amazon, offering a “Buy for Me” option through Amazon or a “Shop Direct” link to other merchants.
Examples of what you can ask Rufus:
- What do I need to host a Frozen-themed birthday party for my daughter?
- Reorder everything we used to make pumpkin pie last week.
- Find me work dresses under $100 that match this photo.
- Compare this laptop with the one already in my cart.
Prices, Deals, And Support Inside The Chat
Rufus now exposes 30 and 90-day price history on products, so you can see at a glance whether a current offer is actually a good value.
Prime members can set price alerts, and the assistant can automatically buy items when they hit a chosen price, using default payment and address details, with a 24-hour cancellation window. Early users are saving around 20 percent on average when they use auto buy.
On top of that, Rufus can filter by price band, surface personalized deals tied to your browsing and wish lists, and track orders.
It also works as a first-line customer service layer, helping with delivery updates, returns, account issues, and escalating to human agents when needed, all within the same chat interface.
Conclusion
Rufus is moving from a simple chatbot to a full shopping copilot, combining generative and agentic AI with memory, visual search, and automated purchasing. Amazon’s own metrics suggest that deeper personalization and better tools are driving higher engagement and conversion across its store.
At the same time, giving an AI assistant long-term account memory and purchase authority makes the trade-offs around data, control, and trust more concrete for everyday shoppers.
Amazon is betting that clear controls, tangible savings, and a smoother journey from idea to checkout will keep most customers happy to keep chatting with Rufus.
📈 Latest AI News
18th November 2025
For the recent AI News, visit our site.
If you liked this article, be sure to follow us on X/Twitter and also LinkedIn for more exclusive content.
