Trainer surrounded by icons floating above woman’s hand
Amazon’s February launch of Rufus, an AI shopping assistant, offers an early glimpse into how artificial intelligence could reshape product discovery and purchase behavior. However, early experiments reveal both the promise and limitations of AI shopping agents, while raising critical questions about how brands should prepare for a future where traditional keyword search may no longer dominate.
For nearly a decade, brands selling on Amazon have followed a relatively linear playbook. Having founded an Amazon-focused marketing agency in 2015, I’ve watched this evolution firsthand. Over that 9-year timeframe, the fundamentals haven’t changed much: optimize for Amazon’s A9 algorithm through strategic keyword use, maintain strong product ratings, build traffic and conversion volume through advertising, and create compelling visual content. While tools and capabilities have become more sophisticated—from enhanced advertising options to advanced content features—the core strategy has remained consistent.
Rufus potentially disrupts this established pattern. Early analysis suggests it considers signals beyond Amazon’s traditional ranking factors, including content from brand websites and broader internet presence.
Brands now have to consider how to optimize their product content, not just for consumers, but also for these large language models. The search terms that consumers may have historically used to find your brand or product won’t matter as much as what the AI model thinks of your brand or product.
The Wider AI Shopping Race
The competition to develop AI shopping assistants extends beyond Amazon. Perplexity’s launch this month of its shopping integration, which enables single-click purchasing within search results, signals growing innovation in this space. This development could pressure other retailers to accelerate their AI initiatives to avoid losing search traffic to third-party platforms.
Amazon’s Rufus AI: Early Limitations vs Future Potential
Reddit user /HomeBrewDude shares how Amazon’s Rufus AI appears to be unsure about its own identity.
Current feedback on Rufus has been mixed, with some users noting its frequently incorrect or unhelpful responses.
However, this mirrors the early days of many transformative technologies. As fellow Forbes contributor Jason Goldberg noted in a recent post about AI shopping agents, “AI shopping assistants aim to bypass [traditional search]
entirely, merging product discovery and purchase into a single, intuitive experience.” While this vision remains largely unrealized, AI shopping assistants represent the potential for a tireless, knowledgeable shopping companion that understands every product detail and customer preference.
This development aligns with predictions I reported last month about the future of e-commerce, where industry experts suggested that traditional search boxes would eventually be replaced by conversational AI agents.
The path to that future, however, may not be as straight as some expect. Technology analyst Benedict Evans, in his 2024 presentation focused exclusively on AI, raises important questions about whether the rapid pace of AI advancement we’ve seen over the past three years can continue. The blistering speed of progress might slow considerably as we tackle increasingly complex challenges.
The Personalization Promise
One area where early experiments with Rufus show promise is in personalization. Analysis suggests that recommendations vary significantly based on users’ historical purchase behavior and typical price points—perhaps the clearest example yet of meaningful personalization in retail search.
Carter Jensen, Senior Manager of Enterprise Marketing Capabilities at General Mills, used his nights and weekends to create a browser extension that runs search queries inside Rufus, then export the results in bulk. Jensen’s early experiments with Rufus reveal surprising patterns in how the AI assistant surfaces products and makes recommendations.
The author’s search for gluten-free dog food options on Amazon’s Rufus AI assistant.
His testing shows that responses vary significantly between users, with evidence suggesting this variation stems from individual shopping patterns and price preferences. For example, when comparing results across different users for the same product categories, Rufus appears to tailor recommendations based on each user’s historical spending thresholds. This is particularly evident in categories like pet food, where price sensitivity varies widely between consumers.
Interestingly, traditional Amazon ranking signals like “Amazon’s Choice” badges don’t seem to carry the same weight in Rufus’s recommendations as they do in standard search results. This suggests that Amazon may be developing a different framework for AI-driven product discovery that relies more heavily on personal shopping history than conventional ranking factors.
Strategic Implications for Brands
While it’s too early to completely rewrite the Amazon playbook, brands should consider several emerging priorities:
- High-Consideration Categories: Brands selling complex or expensive products should pay particular attention. As Jensen notes, while AI assistance may offer limited value for routine $5 purchases, it could dramatically impact categories where consumers invest more time and emotional energy in decision-making—like an $80 bag of dog food where shoppers are “very technical and very emotionally invested.”
- Digital Presence and Content Evolution: Success requires maintaining consistent, high-quality content across all touchpoints, optimized for both traditional search and AI interpretation. This includes considering how AI indexes different content types, from image text to bullet points, and understanding that assistants may pull information from sources beyond retailer platforms.
- Price Positioning: With early signs suggesting strong price sensitivity in AI recommendations, brands need to carefully consider their pricing strategy and how it aligns with target customer segments.
Will Sponsored Ads Be Shown in Amazon Rufus?
An update to Amazon API documentation in November reveals another layer to Rufus’s potential impact: the platform may begin incorporating sponsored ads into Rufus-related placements.
“To help customers discover more products in Amazon’s generative AI-powered shopping assistant, referred to as Rufus, your ads may appear in Rufus-related placements,” says Amazon in its November release notes. “Rufus may generate accompanying text based on the context of the conversation.”
While Amazon notes that campaign reports won’t include Rufus metrics separately, this development raises interesting questions about the future of retail media. How will AI shopping assistants change the traditional advertising model where brands bid on keywords? The inability to track performance specifically within Rufus conversations could also challenge advertisers’ ability to optimize their spending effectively.
Looking Ahead
While the pace of AI advancement remains uncertain, AI shopping assistants like Amazon’s Rufus will inevitably improve. Jensen predicts that by 2025, major retailers could offer AI-powered predictive commerce with dynamic, conversation-driven shopping carts. “The technology is there. And the data is there,” he notes. “It’s just bringing that to life.” For brands, success lies in building flexible foundations that work for both current and future commerce paradigms.