Amazon quietly launched “Rufus Pro” last week — an AI personal shopping assistant powered by a custom version of its Titan language model — and in less than seven days it has already influenced over $340 million in purchase decisions on the platform, according to internal metrics shared with investors at Amazon’s Q1 earnings call. The launch marks a turning point in how AI is reshaping consumer spending: for the first time, a major retailer’s AI assistant isn’t just helping people find products — it’s actively predicting what they’ll want before they’ve thought to search for it.
How AI Shopping Assistants Actually Work
Rufus Pro draws on a customer’s full Amazon purchase history, browsing behaviour, wishlist data, and — for users who opt in — their Alexa voice commands and Prime Video viewing habits to build a continuous behavioural profile. The system uses this profile to generate personalised product recommendations, anticipate repurchase timing (it knows you typically buy laundry detergent every 34 days), and flag deals on items you’ve previously viewed but didn’t buy. In product testing, Amazon found that customers who engaged with Rufus Pro recommendations had a 23% higher average order value and a 31% improvement in purchase satisfaction scores. Google has a competing product deeply integrated into Search and Maps, while Apple is working on a Siri-powered version expected to launch with iOS 20 later this year.
“We are entering a world where your shopping assistant knows your preferences better than you do consciously. The question isn’t whether this technology will dominate retail — it’s whe4her consumers will accept the privacy trade-off it requires.”
— Sucharita Kodali, Vice President and Principal Analyst, Forrester Research
Privacy advocates have raised immediate concerns. Rufus Pro’s opt-in data sharing terms — buried in a 4,200-word update to Amazon’s privacy policy rolled out last month — allow the company to share anonymised “purchase intent signals” with its 40,000-strong third-party seller network, enabling those sellers to adjust pricing and inventory in near real-time. Critics at the Electronic Frontier Foundation argue this creates a surveillance-driven marketplace where sellers can extract maximum value from consumers with a demonstrated willingness to pay. Amazon disputes this characterisation, saying the data-sharing is “aggregate and non-identifiable.”
The competitive landscape is accelerating rapidly. Walmart is testing its own AI shopping assistant — built on Microsoft’s Azure AI platform — across 500 stores and its online platform, with a broader rollout planned for Q3. Meanwhile, Shopify announced a partnership with Anthropic to offer smaller merchants an AI sales assistant powered by Claude, levelling the playing field between independent retailers and the major platforms. The race to own the AI shopping layer of consumer spending is one of the defining competitive battles of 2026.
What This Means For You
If you’re an Amazon Prime member, Rufus Pro is already active on your account. You can review and delete your purchase history data to limit what it knows about you, or opt out of third-party data sharing entirely in your Privacy Settings. Use it as a deal-finding tool for planned purchases rather than a discovery engine for impulse buys — that’s where the real savings are. And be aware that clearing your cookies or using incognito mode before making a major purchase sometimes reveals a lower price than what your logged-in profile is shown.
