This holiday season, major technology companies are heavily promoting new artificial intelligence tools designed to transform how consumers shop online. However, early tests reveal these AI shopping assistants are still in an awkward experimental phase, often delivering generic and uninspired gift recommendations.
The Rush to Automate Holiday Shopping
Retailers and tech giants are racing to capitalize on a predicted shift in consumer behaviour. This comes as United States consumers are expected to spend a record US$253 billion online during the holiday period. The vision, championed by industry prognosticators, is a future where autonomous AI agents handle product research, price comparison, and purchasing, moving beyond traditional search engine queries.
The appeal is obvious. Instead of tediously filtering through millions of products on sites like Amazon or Walmart, a shopper could simply tell a chatbot: “Find me a pair of well-reviewed hiking boots in my size, under US$100, and available for delivery by Friday.” This promises a more intuitive and user-friendly experience. A September survey by Adobe Inc. found that more than one in three U.S. consumers have already used AI tools to assist with online shopping, primarily for product research.
Big Promises, Bathrobe Results
This week, the hype machine was in full swing. Amazon.com Inc. announced new features for its AI shopping assistant, Rufus, calling it a “faster, more useful, state-of-the-art shopping companion.” Google promoted an “agentic checkout feature,” and OpenAI unveiled a free ChatGPT tool that generates personalized gift-buying guides.
Yet, when put to the test, these bots showed their limitations. Bloomberg News recently asked several AI assistants—including Amazon’s Rufus, OpenAI’s ChatGPT, and Walmart Inc.’s Sparky—for a Christmas gift suggestion for mom. The overwhelming top recommendation was a cozy bathrobe. Walmart’s Sparky specifically suggested a pink hooded robe labelled “Mama Bear,” while ChatGPT recommended buying it from Victoria’s Secret. Perplexity Inc.’s AI bot offered another common gift guide item: a US$20 wooden photo frame from Etsy.
The Trillion-Dollar Dream Meets Reality
Consulting firm McKinsey & Co. has forecasted that so-called agentic commerce could become a US$1 trillion business in the U.S. by 2030. For this potential to be realized, technology companies must overcome significant technical hurdles and negotiate crucial partnerships with major retailers and payment processors.
Analysts, however, are tempering expectations for the immediate future. “There are a lot of really big bets being made right now that consumers want to shop differently and that chat is the way they want to start shopping,” said Emily Pfeiffer, an analyst at Forrester Research Inc. “I don’t think this is going to have a huge impact on the way we shop this holiday season.”
While the current output may be underwhelming, there are glimmers of potential. Data from Similarweb Ltd. indicates that shoppers referred to a website after a conversation with ChatGPT tend to be more informed and ready to buy than those arriving from a typical Google search. For now, the journey to a truly intelligent shopping agent continues, one generic bathrobe suggestion at a time.
