Amazon ditches Rufus chatbot, launches Alexa shopping agent in AI strategy pivot

Amazon is axing its Rufus chatbot and making its Alexa assistant the centerpiece of its artificial intelligence shopping strategy.

The company on Wednesday launched Alexa for Shopping, an e-commerce bot that can answer queries and take actions on behalf of users. Amazon said the tool brings together Rufus and Alexa+, and taps users' shopping history and other data to be "the world's best, most personalized AI assistant for shopping."

As part of the move, Amazon is adding Alexa to search results on its store, so if users browse for certain products, a chat window will appear with information and a few recommended items.

A little over two years ago, Amazon unveiled Rufus as a key part of its website and app in an effort to take advantage of the generative AI boom, which was sweeping across the the tech sector and into other parts of the economy. Rufus was described at the time as an "expert shopping assistant," and Amazon continued to expand its capabilities, though it remains in beta.

The stand-alone Rufus chatbot will be discontinued, but Amazon said it will use Rufus' recommendation features and shopping history for certain Alexa for Shopping queries. Users can summon Alexa for Shopping by clicking a cursive A icon on Amazon's website or app, or via Echo Show displays.

Alexa for Shopping turns Amazon's search bar into a Q&A engine, and also lets users compare products side by side, as well as schedule purchases when an item hits a certain price. A Prime membership isn't required to use the tool.

Amazon is evolving its strategy as the e-commerce industry grapples with the rise of AI shopping bots. OpenAI, Google and Perplexity have rolled out research tools and agents over the past year that threaten to disrupt how people shop online. Some of those efforts have stumbled, and it's unclear whether consumers are ready to hand off the task of completing a purchase to bots.

Daniel Rausch, Amazon's top Alexa executive, said the new offering is superior to other AI shopping tools because it has access to valuable data, such as customer reviews and a vast product catalog. It can also reliably tell a user whether a product is in stock, or estimated delivery times, Rausch said.

"As I'm using it, I'm just realizing why other AI efforts have struggled with shopping because it's not just scraping web results and then putting things in a conversation," Rausch said in an interview.

Earlier this year, OpenAI significantly altered its AI shopping plans. The company ended Instant Checkout, a tool that let users check out directly from ChatGPT, in favor of working with retailers to create dedicated apps in its chatbot. OpenAI said at the time that shopping apps would enable users to make purchases "more seamlessly."

Rausch said he wasn't surprised "others have basically had to undo a bunch of features" that were incomplete or disjointed.

"It's just not worth it," he said. "Shopping is not something you do as a side quest."

Amazon has been reluctant to partner with rival AI platforms and open up its site to external shopping agents. CEO Andy Jassy has said the company is "having conversations with" and expects to partner with third-party agents, though Amazon continues to block many bots from accessing its site.

At the same time, it has also launched "Buy for Me," which uses AI to purchase products on a customer's behalf, including products sold on other retailers' websites. The tool sparked backlash from some retailers who said they never opted in to the program.

By inserting Alexa for Shopping into search results, Amazon is taking advantage of valuable real estate for promotion.

The move could prove disruptive to Amazon's millions of third-party sellers, who pay top dollar to promote their listings and rank higher in traditional search results. The ads, which Amazon refers to as sponsored product listings, account for most of the company's advertising revenue.

Alexa for Shopping will feature ads where they're relevant and when they "enhance" the shopping experience, Rausch said, adding that it's not designed to "narrow" search results.

"It's there to, in some cases, expose even more products for customers, depending on where you are in the journey," he said.