Amazon’s new AI shopping assistant “Rufus” : UX review

Evaluating Amazon's Rufus

Amazon is rapidly experimenting with new use cases of Generative AI to help drive more sales and improve the customer experience. One of their main products is Rufus, a conversational shopping assistant embedded directly into the search experience.

What do we at alby think about this experience? Is this something other brands and retailers should adopt? Read below for our review. 

How Amazon's Rufus Works: ChatGPT in the Search Bar

If you type a long-form question into the search bar on the Amazon app, the Rufus shopping assistant will automatically pop up to answer your question and show you relevant products:

Rufus then acts similarly to ChatGPT, allowing shoppers to ask follow up questions and providing more information.

The Good: A New Way to Discover Products

This type of experience enables a new discovery channel that is less transactional than the typical “search, filter, select” shopping flow. Here are some features of Rufus that stand out:

• Good for product bundling: If you want to order all the ingredients you need to make a specific recipe, it can be frustrating to individually search for each item. Rufus could (theoretically) generate a recipe and then make it easy to add all of the products to your cart with one click (a big time saver).
• Unobtrusive (i.e. not an AI chatbot): We appreciate that Amazon did not just “tack-on”' a chatbot pop-up into the shopping experience the way most e-commerce companies do. Instead, Amazon has integrated Rufus into the search experience directly, making it less annoying for the shopper. ‍
• Helpful for selecting complex products: Rufus provides extensive information for niche customer questions. This is helpful for selling complex products where customers have many specific questions they need answered before buying. 
  
  

The Bad: Changing Browsing Behavior is Hard

However, there is one big problem with Amazon’s Rufus: the experience is not intuitive. Very few (if any) shoppers will even know they can use the search bar this way.

Amazon is “overloading” the search bar to try to do two things at once: (1) simple quick searching for products (the old search experience) and (2) long-form complex questions (the new Rufus thing). This is also the strategy that Google is taking with search with their “AI Overview” under search results. Overloading the search bar often confuses shoppers

At alby, we have experimented with this type of experience on various e-commerce websites. We found that less than 0.1% of shoppers type a long-form query into the search bar (even when prompted to do so). Muscle memory is real and 25 years of keyword-based searching has trained shoppers to search in a very specific way.

This means that Amazon’s Rufus will either (1) change shopper behavior to expect search bars to support these types of long-form queries (we are skeptical, but it is possible) or (2) Amazon will abandon this kind of experience for something else (we believe this is more likely).

Amazon’s Other Experiment (Which We Find More Promising)

Amazon has also started to experiment with embedding Rufus directly in product detail pages (PDPs) and automatically predicting an initial set of questions (similar to the alby experience):

Based on our experimentation, we believe that this use case will generate more traction compared to the search bar experience. Embedding Rufus directly into PDPs ensures that customers receive relevant assistance exactly when they need it—while they are considering a purchase. This contextual help can guide customers through their decision-making process without requiring them to leave the page or disrupt their shopping flow.

We believe experiences like this, that don’t “overload” the search bar, are more likely to succeed with shoppers. We have seen 10x more engagement in these types of embedded experiences.

The Key Takeaway

Rufus is a great step forward for using Conversational AI in e-commerce experiences and will certainly impact shopper expectations going forward. However, we are still in the early days of figuring out the right user experience for these AI shopping assistants. While the search bar integration may face challenges due to long ingrained shopper behavior, the embedded experience on product pages shows promise and could become a key feature in enhancing the online journey for Amazon shoppers. 

Brands and retailers should take note of these developments coming from and consider how similar integrations could benefit their customers.