Amazon’s Rufus aims to automate personalized gift shopping with AI

Amazon is trying to remove the hardest part of gift shopping: the blank search bar. With Rufus, its shopping assistant, scheduled actions can now automate repeat buying behavior, including gift picks shaped by prior activity, which means birthdays, holidays and recurring family occasions can move from manual browsing to something closer to a saved routine.

Amazon announced Rufus on February 1, 2024 as a generative-AI-powered conversational shopping experience trained on Amazon’s product catalog and information from across the web. By September 18, 2024, it was available to all U.S. customers in the Amazon Shopping app and on desktop. In November 2024, Amazon explicitly cast Rufus as a holiday-shopping tool that could help people find tailored gifts and inspiration, a signal that the company sees the assistant not just as a search aid, but as a new front door to buying.

The appeal is practical. Amazon says Rufus personalizes shopping using browsing history, wish lists and past purchases, and it can search by activity, event, purpose and other use cases. It can check price history, surface deals, auto-buy items at target prices and shop other merchants. For gift shopping, that can handle the heavy lifting well: finding a second bottle of a fragrance someone already likes, resurfacing a cookbook after a wish-list save, or watching a set of pajamas until the price drops to a number that feels reasonable.

What Rufus still cannot do is the part human gift-givers care about most: judgment. It can remember what was clicked, purchased and saved, but it cannot tell whether a present should feel playful, intimate, restrained or formally generous. That distinction matters in personalized gifting, where a $50 object chosen with precision can feel far more thoughtful than a $500 item that misses the mark. Rufus can narrow the field; it still takes a person to know when the right answer is useful, and when it should be memorable.

The scale behind the assistant suggests Amazon is serious about that shift. Amazon’s cloud blog said more than 250 million customers had used Rufus in the prior year, monthly users were up 140 percent year over year and interactions were up 210 percent year over year. Amazon also said customers who use Rufus during a shopping journey are 60 percent more likely to complete a purchase. Built on AWS infrastructure, including Trainium and Inferentia chips and a custom large language model, Rufus is becoming part of Amazon’s broader AI strategy, and the daily habit it may change most is gift shopping that used to start with tabs and end with fatigue.