Amazon is turning Rufus from a shopping chatbot into a more active retail assistant, with scheduled tasks, price tracking, handwritten list scanning, and tools that can push purchases beyond Amazon's own store.
Amazon's Rufus is moving closer to the kind of AI shopping agent retailers have been talking about for years. The latest update gives the assistant more ways to remember intent, act on it later, and guide customers from product discovery to checkout with less manual searching.
The most important addition is Scheduled Actions. A customer can ask Rufus to recommend healthy snacks for children every month, remind them about birthday gifts, alert them when a favorite author releases a new book, or help plan a household restock. That is a different pattern from ordinary search. Instead of waiting for a shopper to type a fresh query, Amazon is trying to capture intent once and keep acting on it.
According to Amazon, Rufus can already use browsing history, wish lists, and past purchases to personalize recommendations and deal discovery. It can also reorder products tied to previous shopping activity, suggest alternatives when an item is unavailable, and add products to a cart for review before checkout. For customers, that means fewer repeated searches. For Amazon, it means more shopping decisions stay inside its own interface.
The assistant is also becoming more visual. Shoppers can upload a photo and ask Rufus to find a similar product, solve a problem, or identify relevant supplies. For iOS customers, handwritten grocery or holiday lists can now be snapped, transcribed, and turned into cart additions, with Android support expected to follow. That may sound small, but it targets a very ordinary friction point: people still plan purchases in notes, photos, screenshots, and scraps of paper.
Rufus has also added price tools that make it more useful before the moment of purchase. Customers can ask for 30-day or 90-day price history, set a price alert, or ask the assistant to buy an item when it reaches a target price. For Prime members, Amazon says the auto-buy feature can complete the purchase using the customer's default payment method and shipping address, with a notification and a 24-hour cancellation window after the order is placed.
Those features give Amazon a clearer answer to a threat that is forming outside its walls. If shoppers begin asking ChatGPT, Google, Perplexity, or other AI assistants what to buy, Amazon risks losing some of the product discovery it has spent decades owning. Rufus is the countermeasure: an assistant built directly into the store, close to the catalog, reviews, customer data, checkout, and advertising engine.
Amazon says more than 250 million customers have used Rufus this year, with monthly average users up 149% and interactions up 210% over the past year. The company also says customers who use Rufus while shopping are more than 60% more likely to make a purchase during that shopping trip. Those figures do not prove that Rufus is the only reason people buy, but they explain why Amazon is pushing the assistant deeper into the shopping flow.
Consumer Shift
AI shopping is beginning to change from a question-and-answer experience into something more persistent. A customer does not just ask what shoes to buy today. They can ask an assistant to watch prices, remember a preference, prepare for an occasion, or surface the right replacement when a regular item runs out.
That shift matters because online retail has long depended on repeated intent signals: search terms, product page visits, cart additions, and ad clicks. Scheduled Actions compress that journey. If Rufus already knows a shopper wants birthday ideas two weeks before an event, or cleaning supplies every month, the next purchase may begin before the shopper actively returns to search.
This is where Amazon's scale becomes hard to match. Rufus is not starting with a blank profile. It sits on top of a customer's Amazon history, including purchases, browsing behavior, wish lists, cart activity, and conversations with the assistant. Amazon has also said Rufus will gain memory across more of its digital services, including Kindle, Prime Video, and Audible, which could make recommendations feel more personal over time.
There is a tradeoff for shoppers. The more useful Rufus becomes, the more it depends on Amazon interpreting personal behavior correctly. Bad recommendations, sponsored bias, or overconfident answers could quickly undermine trust. AI shopping only works if the customer believes the assistant is saving time rather than steering them toward whatever Amazon most wants to sell.
Enterprise Merchant Tools
For merchants, Rufus changes the importance of product data. A listing that works for human browsing may not be enough when an AI assistant is deciding which products fit a scheduled task, a handwritten list, or a personalized deal query. Attributes, availability, pricing, reviews, delivery speed, and product descriptions all become inputs in an automated recommendation system.
That creates pressure on brands to make their catalog data cleaner and more machine-readable. The old playbook focused heavily on keywords, ad bids, images, and conversion rates. Those still matter. But if Rufus is matching a shopper's saved preferences against product attributes, incomplete listings may lose visibility before the customer ever sees a search results page.
Amazon is also widening Rufus beyond its own shelves through Shop Direct and Buy For Me. Shop Direct can show products from other merchants when Amazon does not sell them, while Buy For Me can let Amazon complete selected purchases on a merchant's site using the customer's Amazon payment and shipping details. TechCrunch reported in March that Amazon expanded Shop Direct by supporting third-party product feeds from providers including Feedonomics, Salsify, and CedCommerce.
That move is strategically important. It lets Amazon remain the starting point even when the product is not on Amazon. For independent merchants, it may bring exposure. For Amazon, it keeps the assistant in the middle of the transaction, preserving the customer relationship and the data trail around what shoppers wanted, compared, and bought.
Advertising is the next question. Reports have already pointed to Amazon exploring ads inside Rufus, and that would fit the company's broader retail media business. But AI answers are a more delicate surface than search results. If customers start to feel that Rufus recommendations are paid placements dressed up as advice, the assistant could lose the trust it needs to become a daily shopping tool.
What To Watch Next
The broader direction is clear. Amazon wants Rufus to become the default layer between intent and purchase: the place where shoppers ask, compare, plan, reorder, and eventually delegate more of the work. Scheduled Actions make that ambition more concrete because they turn future intent into something Amazon can act on without waiting for a new search.
The practical test will be reliability. Rufus has to recommend the right products, respect budgets, explain choices clearly, and give customers enough control before money changes hands. If it does that well, Amazon strengthens its position in e-commerce at the exact moment AI assistants are trying to move shopping away from traditional search. If it stumbles, shoppers will treat it like another feature to ignore.
For now, Rufus shows where online retail is heading. The storefront is becoming conversational, the cart is becoming more automated, and the next battle in e-commerce may be less about who has the biggest catalog and more about whose assistant customers trust to choose from it.
Also read: Meta's AI agents now run entire ad campaigns on Instagram and Facebook • Meta's AI bill comes due: 700 Irish workers at content contractor Covalen face the axe • Archbald's council implodes as six AI data centers threaten to swallow a Pennsylvania town
